Full text of "James Arthur lecture on the evolution of the human brain"

See other formats

S8MflBnmffifflB8ft

BBjgwawiri™

WWII IIWM fflffl ]

.SmHI

MHWI1

•PTOMI tSariSa

jraalga Mlfflliffllfl B

BWJ11CKIII11 1 818I& W

xiUjlHWIjaa

HHflBHl

38f

Hi.

LIBRARY OF THE
/$- FOR THE ^

^ PEOPLE ^

< FOR ^*

EDVCATION O

<S> SCIENCE ^<

1.4

JAMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

1964

PROBLEMS OUTSTANDING IN THE
EVOLUTION OF BRAIN FUNCTION

ROGER W. SPERRY

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1964

355

1869
THE LIBRARY

JAMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

1964

PROBLEMS OUTSTANDING IN THE
EVOLUTION OF BRAIN FUNCTION

ROGER W. SPERRY

Hixon Professor of Psychobiology

California Institute of Technology

Pasadena, California

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1964

JAMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

Frederick Tilney, The Brain in Relation to Behavior; March 15, 1932

C. Judson Herrick, Brains as Instruments of Biological Values; April 6, 1933

D. M. S. Watson, The Story of Fossil Brains from Fish to Man; April 24, 1934

C. U. Ariens Kappers, Structural Principles in the Nervous System; The Develop-
ment of the Forebrain in Animals and Prehistoric Human Races; April 25,
1935
Samuel T. Orton, The Language Area of the Human Brain and Some of its Dis-
orders; May 15, 1936
R W. Gerard, Dynamic Neural Patterns; April 15, 1937
Franz Weidenreich, The Phylogenetic Development of the Hominid Brain and its

Connection with the Transformation of the Skull; May 5, 1938
G. Kingsley Noble, The Neural Basis of Social Behavior of Vertebrates; May 11,

1939
John F. Fulton, A Functional Approach to the Evolution of the Primate Brain;

May 2, 1940
Frank A. Beach, Central Nervous Mechanisms Involved in the Reproductive Be-
havior of Vertebrates; May 8, 1941
George Pinkley, A History of the Human Brain; May 14, 1942
James W. Papez, Ancient Landmarks of the Human Brain and Their Origin;

May 27, 1943
James Howard McGregor, The Brain of Primates; May 11, 1944
K. S. Lashley, Neural Correlates of Intellect; April 30, 1945
Warren S. McCulloch, Finality and Form in Nervous Activity; May 2, 1946
S. R. Detwiler, Structure-Function Correlations in the Developing Nervous Sys-
tem as Studied by Experimental Methods; May 8, 1947
Tilly Edinger, The Evolution of the Brain; May 20, 1948
Donald O. Hebb, Evolution of Thought and Emotion; April 20, 1949
Ward Campbell Halstead, Brain and Intelligence; April 26, 1950
Harry F. Harlow, The Brain and Learned Behavior; May 10, 1951
Clinton N. Woolsey, Sensory and Motor Systems of the Cerebral Cortex; May 7,

1952
Alfred S. Romer, Brain Evolution in the Light of Vertebrate History; May 21,

1953
Horace W. Magoun, Regulatory Functions of the Brain Stem; May 5, 1954
Fred A. Mettler, Culture and the Structural Evolution of the Neural System;

April 21, 1955
Pinckney J. Harman, Paleoneurologic, Neoneurologic, and Ontogenetic Aspects

of Brain Phytogeny; April 26, 1956
Davenport Hooker, Evidence of Prenatal Function of the Central Nervous System

in Man; April 25, 1957
David P. C. Lloyd, The Discrete and the Diffuse in Nervous Action; May 8, 1958
Charles R. Noback, The Heritage of the Human Brain; May 6, 1959
Ernst Scharrer, Brain Function and the Evolution of Cerebral Vascularization;

May 26, 1960
Paul I. Yakovlev, Brain, Body and Behavior. Stereodynamic Organization of the
Brain and of the Motility-Experience in Man Envisaged as a Biological
Action System; May 16, 1961
H. K. Hartline, Principles of Neural Interaction in the Retina; May 29, 1962
Harry Grundfest, Specialization and Evolution of Bioelectric Activity; May 28,

1963
Roger W. Sperry, Problems Outstanding in the Evolution of Brain Function;
June 3, 1964

PROBLEMS OUTSTANDING IN THE
EVOLUTION OF BRAIN FUNCTION

Having been for some time one of those "card-carrying
members" of the American Museum and being indebted to
the Museum on certain other counts over the years, I much
appreciate the invitation to give the 1964 James Arthur
Lecture. I have been forewarned that many in the audience
will not be particularly carried away by the "latest technical
advances." I have also been forewarned that another 50 per
cent are apt to be rather bored by anything else. The motley
mixture of material with which I have tried to balance the
diverse interests will, I fear, strain even so broad an encom-
passing theme as that of evolution. As indicated in the title,
we shall be concerned more with the functional than with
the morphological properties of the brain, and more with
remaining unsolved problems than with the solid progress
over which we already can beat our chests.

I wish to skip the beginning steps in the evolution of the
human brain and pick up the story at about the culmination
of the latter half of the age of hydrogen gas. In such a way
we can bypass what is by far the most difficult of all the
unsolved problems in brain evolution, namely, how, when,
and where did the hydrogen age and the whole business
start? This problem we can leave to the proponents of the
"steady state," the "periodic pulsation," and the "big bang,"
at least until someone comes along with a more credible
interpretation of the meaning of the red shift.

We can skip quickly also through those early periods
when, first, electrons and protons were being used to build
bigger and better atoms, and then the atoms to make bigger
and finer molecules, and then these in turn were being com-
pounded into giant and replicating molecules and self-

organizing molecular complexes and eventually that elabo-
rate unit, the living cell.

We need pause here only to note for future reference that
evolution keeps complicating the universe by adding new
phenomena that have new properties and new forces and
that are regulated by new scientific principles and new
scientific laws — all for future scientists in their respective
disciplines to discover and formulate. Note also that the old
simple laws and primeval forces of the hydrogen age never
get lost or cancelled in the process of compounding the
compounds. They do, however, get superseded, over-
whelmed, and outclassed by the higher-level forces as these
successively appear at the atomic, the molecular, and the
cellular and higher levels.

We can turn now to what is probably the "most unan-
swered" problem in brain evolution. We encounter it a bit
later on, presumably after organisms with nerve nets and
brains have entered the picture. I refer, as you probably
guess, to the first appearance of that most important of all
brain properties and certainly the most precious, conscious
awareness. (I hope that it is safe to assume that, since "mind"
and "consciousness" have made a comeback in recent years
and have become respectable terms again in the Boston area,
it is permissible to mention them here also.)

In any case, the fossil record notwithstanding, there
seems to be good reason to regard the evolutionary debut
of consciousness as very possibly the most critical step in
the whole of evolution. Before this, the entire cosmic
process, we are told, was only, as someone has phrased it,
"a play before empty benches" — colorless and silent at that,
because, according to our best physics, before brains there
was no color and no sound in the universe, nor was there
any flavor or aroma and probably rather little sense and no
feeling or emotion.

All of these can now be generated by the surgeon's elec-

trode tip applied to the proper region of the exposed con-
scious brain. They can be triggered also, of course, by the
proper external stimuli, but also, more interestingly, by
centrally initiated dream states, illusionogenic and hal-
lucinogenic agents, but always and only within and by a
brain. There probably is no more important quest in all
science than the attempt to understand those very particular
events in evolution by which brains worked out that special
trick that has enabled them to add to the cosmic scheme of
things: color, sound, pain, pleasure, and all the other facets
of mental experience.

In searching brains for clues to the critical features that
might be responsible, I have never myself been inclined to
focus on the electrons, protons, or neutrons of the brain, or
on its atoms. And, with all due respect to biochemistry and
the N.R.P., I have not been inclined to look particularly at
the little molecules of the brain or even at its big macro-
molecules in this connection. It has always seemed rather
improbable that even a whole brain cell has what it takes
to sense, to perceive, to feel, or to think on its own. The
"search for psyche," in our own case at least, has been di-
rected mainly at higher-level configurations of the brain,
such as specialized circuit systems, and not just any juicy
central nerve network that happens to be complex and
teeming with electrical excitations. I have been inclined to
look rather at circuits specifically designed for the express
job of producing effects like pain, or High C or blue-yellow
— circuits of the kind that one finds above a high transec-
tion of the spinal cord but not below, circuits with something
that may well be present in the tiny pinhead dimensions of
the midbrain of the color-perceiving goldfish but lacking in
the massive spinal-cord tissue of the ox, circuits that are
profoundly affected by certain lesions of the midbrain and
thalamus but little altered by complete absence of the entire
human cerebellum. Were it actually to come to laying our

money on the line, I should probably bet, first choice, on
still larger cerebral configurations, configurations that in-
clude the combined effect of both (a) the specialized circuit
systems such as the foregoing plus (b) a background of
cerebral activity of the alert, waking type. Take away either
the specific circuit, or the background, or the orderly activ-
ity from either one, and the conscious effect is gone.

In this day of information explosion, these matters are
not so much of the "ivory tower" as they used to be: To
the engineer who comes around from Industrial Associates
with dollars and cents in his eye and company competition
in his heart the possibility is of more than theoretical inter-
est that conscious awareness may be something that is not
necessarily tied to living hardware, that it could prove to be
an emergent, over-all circuit property that might, in theory,
be borrowed and, given sufficient acreage, perhaps copied
some day in order to incorporate pain and pleasure, sensa-
tions and percepts, into the rapidly evolving circuitry of
computer intellect. When the aim is to build into your
circuit systems some kind of negative and positive reinforce-
ment, then pain and pleasure are about the best kind. And
eager young theoreticians from the NASA committee or
from radio astronomy already want a more educated guess
about the possibility of encountering on other globes other
minds with perhaps totally different dimensions of conscious
awareness, and if not, why not? Then there are more immi-
nent, practical matters such as the need, in view of certain
other explosions we face, to be able to pinpoint the first
appearance of consciousness in embryonic development and
chart its subsequent growth and maturation.

Unless you are among those who still believe that value
judgments lie outside the realm of science, you may prob-
ably agree that a few reliable answers in these general areas
and their implications could shake considerably the going
value systems of our whole culture.

We shift now to certain lesser and subsidiary problems,
but problems more approachable in research. Unlike the
situation 25 years ago, most of us today are quite ready to
talk about the evolution and inheritance not only of brain
morphology but also of brain function, including general
behavior and specific behavior traits. Earlier renunciation
of the whole instinct concept in the animal kingdom gen-
erally stemmed in large part from our inability to imagine any
growth mechanisms sufficiently precise and elaborate even
to begin the fabrication of the complex nerve networks of
behavior. This outlook was supported in the analytic studies
of nerve growth all through the 1920's and 1930's which indi-
cated that nerve fibers grow and connect in a random, dif-
fuse, and non-selective manner governed almost entirely by
indifferent, mechanical factors.

Today the situation is entirely changed. The supposed
limitations in the machinery of nerve growth are largely
removed in the new insight that we have obtained in recent
years into the way in which the complicated nerve-fiber
circuits of the brain grow, assemble, and organize them-
selves in a most detailed fashion through the use of intricate
chemical codes under genetic control (Sperry, 1950a,
1950b, 1951, 1958, 1961, 1962, 1963). The new outlook
holds that the cells of the brain are labeled early in develop-
ment with individual identification tags, chemical in nature,
whereby the billions of brain cells can thereafter be recog-
nized and distinguished, one from another. These chemical
differentials are extended into the fibers of the maturing
brain cells as these begin to grow outward, in some cases
over rather long distances, to lay down the complicated
central communication lines. It appears from our latest evi-
dence that the growing fibers select and follow specific pre-
scribed pathways, all well marked by chemical guideposts
that direct the fiber tips to their proper connection sites.
After reaching their correct synaptic zones, the fibers then

link up selectively among the local population with only
those neurons to which they find themselves specifically
attracted and constitutionally matched by inherent chemical
affinities.

The current scheme now gives us a general working pic-
ture of how it is possible, in principle at least, for behavioral
nerve nets of the most complex and precise sorts to be built
into the brain in advance without benefit of experience.
Being under genetic control, these growth mechanisms are
of course inheritable and subject to evolutionary develop-
ment. The same is true of the differential endogenous
physiological properties of the individual cell units in these
networks which, along with the morphological intercon-
nections, are of critical importance in the shaping of be-
havior patterns. We have at present only the general outlines
and general principles of the developmental picture; much
of the detail has yet to be worked out. Also, the underlying
chemistry of the demonstrated selectivity in nerve growth,
as well as the molecular basis of the morphogenetic gradients
involved, and of all the rest of the chemical "I. D. Card"
concept remains a wide-open field that so far has been
virtually untouched.

In connection with this emphasis on the "inherited" in
brain organization, one may well question the extent to
which the observed inbuilt order in the anatomical structure
necessarily conditions functional performance and behavior.
Some years ago, when we subscribed to the doctrine of an
almost omnipotent adaptation capacity in the central ner-
vous system and to functional equipotentiality of cortical
areas and to the functional interchangeability among nerve
connections in general (Sperry, 1958), Karl Lashley sur-
mised that if it were feasible, a surgical rotation through
180 degrees of the cortical brain center for vision would
probably not much disturb visual perception. Rotation of
the brain center was not feasible, but it was possible to

rotate the eyes surgically through 180 degrees in a number
of the lower vertebrates and also to invert the eyeball, by
transplantation from one to the other orbit, on the up-down
or on the front-back axis and also to cross-connect the right
and left eyes to the wrong side of the brain (Sperry, 1950a) .
All these and different combinations thereof were found
to produce very profound disturbances of visual perception
that were correlated directly in each case with the geometry
of the sensory disarrangement. The animals, after recovery
from the surgery, responded thereafter as if everything were
to them upside down and backward, or reversed from left
to right, and so on. Contrary to earlier suppositions regard-
ing the dynamics of perception and cortical organization,
it appeared that visual perception was very closely tied
indeed to the underlying inherited structure of the neural
machinery.

We inferred further from nerve-lesion experiments
(Sperry, 1950b) on the illusory spinning effects produced
by these visual inversions that the inbuilt machinery of
perception must include also certain additional central
mechanisms by which an animal is able to distinguish those
sensory changes produced by its own movement from those
originating outside. The perceptual constancy of an environ-
ment in which an animal is moving, for example, or of an
environment that it is exploring by eye, head, or hand move-
ments, would seem to require that, for every movement
made, the brain must fire "corollary discharges" into the
perceptual centers involved. These anticipate the displace-
ment effect and act as a kind of correction or stabilizing
factor. These centrally launched discharges must be differ-
entially gauged for the direction, speed, and distance of
each move. Along with the dynamic schema for body posi-
tion which the brain must carry at all times, these postulated
discharges conditioning perceptual expectancy at every
move would appear to be a very important feature of the

unknown brain code for perception. The consistent appear-
ance of the spontaneous optokinetic reaction of inverted
vision in fishes, salamanders, and toads would indicate that
the underlying mechanism is basic and must have evolved
very early.

Since the representation of movement at higher cortical
levels generally seems to be more in terms of the perceptual
expectancy of the end effect of the movement than in terms
of the actual motor patterns required to mediate the move-
ment, the postulated "corollary discharges" of perceptual
constancy may not involve so much of an additional load,
in terms of data processing, as might at first appear.

We are ready now for that old question: How much of
brain organization and behavior should we blame or credit
to inheritance and how much to learning and experience?
As far as we can see now, it seems fair to say that all that
central nervous organization that is illustrated and described
in the voluminous textbooks, treatises, and professional
journals of neuro-anatomy, that is, all the species-constant
patterning of brain structure, the micro-architecture as well
as the gross morphology that has so far been demonstrated
anatomically, seems to be attributable to inheritance. An-
other way of saying the same thing is that no one has yet
succeeded in demonstrating anatomically a single fiber or
fiber connection that could be said with assurance to have
been implanted by learning. In this same connection, it is
entirely conceivable (though not particularly indicated)
that the remodeling effects left in the brain by learning and
experience do not involve the addition or subtraction of
any actual fibers or fiber connections but involve only
physiological, perhaps membrane, changes that effect con-
ductance or resistance to impulse transmission, or both, all
within the existing ontogenetically determined networks.

The foregoing picture leaves plenty of room for learning
and for the combined effects of learning plus maturation

during that prolonged period in human childhood when
these two factors overlap. Nevertheless the present picture
represents a very considerable shift of opinion over the past
two decades in the direction of inheritance.

Some of you may find certain aspects here a bit difficult
to reconcile with other inferences drawn in recent years
from a series of sensory deprivation studies on mammals in
which cats, monkeys, chimpanzees, and other animals have
been raised in the dark or with translucent eye caps or in
harnesses or holders of various sorts and in which, as a
result of the various kinds of deprivation of experience in
their early development the animals came to show subse-
quent deficits, moderate to severe, in their perceptual or
motor capacities. The tendency to interpret these findings,
along with those from human cataract cases, as evidence
of the importance of early learning and experience in
shaping the integrative organization of the brain we have
long felt to have been overdone (Sperry, 1950a, 1962).
In nearly all cases the findings could be equally well
explained on the assumption that the effect of function is
simply to maintain, or to prevent the loss of, neural organ-
ization already taken care of by growth. What the results
have come to show in many of these studies is that certain
of the newly formed neuronal elements, if abnormally
deprived of adequate stimulation, undergo an atrophy of
disuse. In much the same way cells of the skeletal muscles
differentiate in development to the point at which they are
contractile and ready to function, but then they too atrophy
and degenerate if not activated. This basic developmental
"use-dependent" property in maturing neurons, or even
some evolutionary derivative of it, applied farther centrally
beyond the sensory paths amid more diffuse growth pres-
sures, especially among cortical association units, could,
however, have true patterning effects and become a definitely
positive factor in learning and imprinting.

We have been approaching very closely here the general
problem of memory. Among brain functions, memory cer-
tainly rates as one of the prime "problems outstanding."
Whatever the nature of the neural mechanism underlying
memory, it seems to have appeared quite early in evolution.
(Some writers say that even flatworms have memory!) We
are frequently impressed in our own work with learning
and memory in cats, and even in fishes, with the fact that
their simple memories, once implanted, seem to be strong
and lasting. With respect to memory, then, what separates
the men from the animals is very likely not so much the
nature of the neural trace mechanism as the volume and
the kind of information handled. The problems that relate
to the translation and coding of mental experience, first
into the dynamics of the brain process, and then into the
static, frozen, permanent trace or engram system, pose the
more formidable aspects of the memory problem.

Fundamental to these memory questions, as also to the
problems of perception, volition, learning, motivation, and
most of the higher activities of the nervous system, is that
big central unknown that most of us working on the higher
properties of the brain keep tangling with and coming back
to. You may find it referred to variously as: the "brain
code," or the "cerebral correlates of mental experience," or
the "unknown dynamics of cerebral organization," or the
"intermediary language of the cerebral hemispheres," or,
in some contexts, just the "black box." Thus far we lack
even a reasonable hypothesis regarding the key variables
in the brain events that correlate with even the simplest of
mental activities, such as the elementary sensations or the
simple volitional twitch of one's little finger.

In our own efforts to help to chip away at this central
problem of the language of the hemispheres, we have been
trying for some 10 years first to divide the problem in half by
splitting the brain down the middle before we start to study

it. (Many times we wonder if the end effect of this split
brain approach is not so much to halve our problems as
it is to double them.) At any rate, the brain-bisection
studies leave us with a strong suspicion that evolution may
have saddled us all with a great deal of unnecessary dupli-
cation, both in structure and in the function of the higher
brain centers.

Space in the intracranial regions is tight, and one won-
ders if this premium item could not have been utilized for
better things than the kind of right-left duplication that
now prevails. Evolution, of course, has made notable errors
in the past, and one suspects that in the elaboration of the
higher brain centers evolutionary progress is more encum-
bered than aided by the bilateralized scheme which, of
course, is very deeply entrenched in the mechanisms of
development and also in the basic wiring plan of the lower
nerve centers.

Do we really need two brain centers, for example, to tell
us that our blood sugar is down or our blood pressure is
up, or that we are too hot or too cold, and so on? Is it
necessary to have a right and also a left brain center to let
us know that we are sleepy or angry, sad or exuberant, or
that what we smell is Arpege or what we taste is salty or
that what we hear is voices, and so on and on and on?
Surely most of us could manage to get along very well with
only one cerebral anxiety mechanism, preferably in the
minor hemisphere.

Emotion, personality, intellect, and language, among
other brain business, would seem by nature to be quite
manageable through a single unified set of brain controls.
Indeed, the early loss of one entire hemisphere in the cat,
monkey, and even in man causes amazingly little deficit
in the higher cerebral activities in general.

With the existing cerebral system, most memories as well
have to be laid down twice — one engram for the left

hemisphere and another engram copy for the right hemi-
sphere. The amount of information stored in memory in
a mammalian brain is a remarkable thing in itself; to have
to double it all for the second hemisphere would seem in
many ways a bit wasteful. It is doubtful that all this
redundancy has had any direct survival value (unless evolu-
tion could have foreseen that neurologists would be opening
and closing the cranium to produce brain lesions under
careful aseptic conditions that permit survival).

In the human brain, of course, we begin to see definite
evidence of a belated tendency in evolution to try to cir-
cumvent some of the duplication difficulties. A de-duplica-
tion trend is seen particularly in the lateralization of speech
and writing within the single dominant hemisphere in the
majority of persons. Speech, incidentally, is another essen-
tially symmetrical activity for which a double right and
left control is quite unnecessary, even at the lower levels
of the motor hierarchy. When the brain does try in some
individuals to set up two central administrations for speech,
one in each hemisphere, the result tends to make for
trouble, like stammering and a variety of other language
difficulties.

The fact that the corpus callosum interconnecting the
two cerebral hemispheres is so very large and the functional
damage produced by its surgical section is so very minor
in most ordinary activities seems to be explainable in part
by the fact that the great cerebral commissure is a system
for cross communication between two entities that to a
large extent are each completely equipped and functionally
self-sufficient. The corpus callosum appears late in evolu-
tion, being essentially a mammalian structure, and its
development is closely correlated with the evolutionary
elaboration of the neocortex of the mammalian cerebral
hemispheres.

Accordingly it is not surprising that it is in the human

brain, and particularly in connection with speech, that the
functional effects produced by surgical disconnection of the
two cerebral hemispheres become most conspicuous. Dur-
ing the past two years we have had an opportunity to test
and to study two patients, formerly unmanageable epilep-
tics, who have had their right and left hemispheres discon-
nected by complete section of the corpus callosum, plus
the anterior commissure, plus the hippocampal commissure,
plus the massa intermedia, in what is perhaps the most
radical surgical approach to epilepsy thus far undertaken.
The surgery was done by Drs. Philip J. Vogel and Joseph
E. Bogen (see Bogen and Vogel, 1962) of Los Angeles. 1
It seemed a reasonable hope, in advance, that such surgery
might help to restrict the seizures to one hemisphere and
hence to one side of the body, and possibly to the distal
portions of arm and leg, since voluntary control of both
sides of the head, neck, and trunk tends to be represented
in both hemispheres. In our colony of split-brain monkeys
that have had similar surgery, we not uncommonly see
epileptic-like seizures, especially during the early weeks
after brain operations, and these seizures show a definite
tendency to center in the distal extremities of the arm and
leg and to be restricted to one side. It also seemed reason-
able that this surgery might help the patients to retain
consciousness in one hemisphere during an attack, if not
throughout, at least during the early stages, and thereby
give them a chance to do things that might help to break

1 The surgical treatment of these cases was undertaken at the suggestion of
Dr. Bogen after extensive consultations on all aspects. The surgery was per-
formed by Dr. Vogel, assisted by Dr. Bogen and other staff members at the
Loma Linda Neurosurgical Unit, White Memorial Hospital. Most of the tests
reported here were planned and administered by Michael S. Gazzaniga of our
laboratory, with the writer collaborating on a general advisory basis.

The original work included here was supported by grants from the National
Institutes of Health (MH-03372, S-F1-MH-18,080) and by the Frank P. Hixon
Fund of the California Institute of Technology.

or control the seizures, or at a rninimum to allow time in
which to undertake protective measures before the second
side became involved. It was further hoped that such
surgery might reduce the severity of the attacks by the
elimination of a very powerful avenue for the right-left
mutual reinforcement of the seizures during the generalized
phase, especially during status epilepticus, which was of
major concern in both the cases cited above.

Judged from earlier reports of the cutting of the corpus
callosum and from the behavior of dozens of monkeys that
we have observed in the laboratory with exactly the same
surgery (Sperry, 1961, 1964), it seemed unlikely that this
kind of surgery would produce any severe handicap, or
surely none so bad as certain other forms of psychosurgery
that have been used on a much more extensive scale. That
the surgery might decrease the incidence of the seizures to
the point of virtually eliminating them (as it seems to have
done so far in both cases) was unexpected; our fingers
remain very much crossed on this latter point.

Everything that we have seen so far indicates that the
surgery has left each of these people with two separate
minds, i.e., with two separate spheres of consciousness
(Gazzaniga, Bogen, and Sperry, 1962, 1963). What is
experienced in the right hemisphere seems to be entirely
outside the realm of awareness of the left hemisphere. This
mental duplicity has been demonstrated in regard to percep-
tion, cognition, learning, and memory. One of the hemi-
spheres, the left, dominant, or major hemisphere, has speech
and is normally talkative and conversant. The other mind
of the minor hemisphere, however, is mute or dumb, being
able to express itself only through non-verbal reactions;
hence mental duplicity in these people following the sur-
gery, but no double talk.

Fortunately, from the patients' standpoint, the functional
separation of the two hemispheres is counteracted by a large

number of unifying factors that tend to keep the discon-
nected hemispheres doing very much the same thing from
one part of the day to the next. Ordinarily, a large common
denominator of similar activity is going in each. When we
deliberately induce different activities in right and left hem-
ispheres in our testing procedures, however, it then appears
that each hemisphere is quite oblivious to the experiences
of the other, regardless of whether the going activities
match or not.

This is illustrated in many ways: For example, the
subject may be blindfolded, and a familiar object such as
a pencil, a cigaret, a comb, or a coin is placed in the left
hand. Under these conditions, the mute hemisphere con-
nected to the left hand, feeling the object, perceives and
appears to know quite well what the object is. It can
manipulate it correctly; it can demonstrate how the object
is supposed to be used; and it can remember the object
and go out and retrieve it with the same hand from among
an array of other objects. While all this is going on, the
other hemisphere has no conception of what the object is
and says so. If pressed for an answer, the speech hemisphere
can only resort to the wildest of guesses. So the situation
remains just so long as the blindfold is kept in place and
other avenues of sensory input from the object to the talk-
ing hemisphere are blocked. But let the right hand cross
over and touch the test object in the left hand; or let the
object itself touch the face or head as in the use of a comb,
a cigaret, or glasses; or let the object make some give-away
sound, such as the jingle of a key case, then immediately
the speech hemisphere produces the correct answer.

The same kind of right-left mental separation is seen in
tests involving vision. Recall that the right half of the visual
field and the right hand are represented together in the
left hemisphere and vice versa. Visual stimuli such as pic-
tures, words, numbers, and geometric forms flashed on a

screen directly in front of the subject and to the right side
of a central fixation point are all described and reported
correctly with no special difficulty. On the other hand
similar material flashed to the left half of the visual field is
completely lost to the talking hemisphere. Stimuli flashed
to one half field seem to have no influence whatever, in
tests to date, on the perception and interpretation of stimuli
presented to the other half field.

Note in passing that these disconnection effects do not
show up readily in ordinary behavior. They must be demon-
strated by the flashing of the visual material fast enough so
that eye movements cannot be used to sneak the answers
into the wrong hemisphere, or in the testing of right and
left hands vision must be excluded with a blindfold, audi-
tory cues eliminated, and the hands kept from crossing,
and so on. One of the patients, a 30-year old housewife
with two children, goes to market, runs the house, cooks
the meals, watches television, and goes out to complete,
three-hour shows at the drive-in theater, all without com-
plaining of any particular splitting or doubling in her per-
ceptual experience. Her family believes that she still does
not have so much initiative as formerly in her houseclean-
ing, in which she was meticulous, and that her orientation
is not so good, for example, she does not find her way back
to the car at the drive-in theater as readily as she formerly
could. In the early months after surgery there appeared to
be definite difficulty with memory. By now, some eight
months later, there seems to be much improvement in this
regard, though not complete recovery. Involvement of the
fornix would have to be ruled out before effects like the
latter can be ascribed to the commissurotomy.

In the visual tests again, one finds plenty of evidence that
the minor, dumb, or mute hemisphere really does perceive
and comprehend, even though it cannot express verbally
what it sees and thinks. It can point out with the left hand

a matching picture from among many others that have been
flashed to the left field, or it can point to a corresponding
object that was pictured in the left-field screen. It can also
pick out the correct written name of an object that it has
seen flashed on the screen, or vice versa. In other words,
Gazzaniga's more recent results show that the dumb left
hemisphere in the second patient is not exactly stupid or
illiterate; it reads a word such as "cup," "fork," or "apple"
flashed to the left field and then picks out the corresponding
object with the left hand. While the left hand and its hemi-
sphere are thus performing correctly, however, the other
hemisphere, again, has no idea at all which object or which
picture or which name is the correct one and makes this
clear through its verbal as well as other responses. You
regularly have to convince the talking hemisphere to keep
quiet and to let the left hand go ahead on its own, in which
case it will usually pick out the correct answer.

These minor differences of opinion between the right
and left hemispheres are seen rather commonly in testing
situations. For example, the left hand is allowed to feel and
to manipulate, say, a toothbrush under the table or out of
sight behind a screen. Then a series of five to 10 cards are
laid out with names on them such as "ring," "key," "fork."
When asked, the subject may tell you that what she felt in
the left hand was a "ring." However, when instructed to
point with the left hand, the speechless hemisphere delib-
erately ignores the erroneous opinions of its better half and
goes ahead independently to point out the correct answer,
in this case the card with the word "toothbrush."

As far as we can see, about the only avenue remaining
for direct communication between mind-right and mind-left
is that of extrasensory perception. If any two minds should
be able to tune in on each other, one might expect these two
to be able to do so, but thus far no evidence of such effects
is apparent in the test performances.

The conscious awareness of the minor hemisphere pro-
duced by this vertical splitting of the brain often seems so
remote to the conversant hemisphere as to be comparable
perhaps to that produced by a spinal transection. To go
back here to some of the issues on which we started, one
wonders if we can really rule out, as I implied above, the
alternative contention of those who maintain that spinal
cords, loaves of bread, and even single molecules have a
kind of consciousness. Either way, the inferences to be
drawn regarding the evolution and elaboration of conscious-
ness for most practical purposes remain much the same.

We are often asked if each of the disconnected hemi-
spheres must not also have a will of its own and if the two
do not then get into conflict with each other. In the first
half year after surgery, particularly with the first patient,
we got reports suggesting something of the kind. For ex-
ample, while the patient was dressing and trying to pull on
his trousers, the left hand started to work against the right,
pulling them off again. Or, the left hand, after just helping
to tie the belt of his robe, went ahead on its own to untie
the completed knot, whereupon the right hand would have
to supervene again to get it retied. The patient and his wife
used to refer to the "sinister left hand" which sometimes
tried to push the wife away aggressively at the same time
that the hemisphere of the right hand was trying to get her
to come and help him with something. These antagonistic
movements of right and left hands are fairly well restricted
to situations in which the reactions of left and right hand
are easily made from the same common supporting posture
of body and shoulders. Generally speaking, there are so
many unifying factors in the situation and functional har-
mony is so strongly built into the undivided brain stem and
spinal networks, by express design, that one sees little overt
expression or overflow into action, at least, of conflicting
will power.

This matter of having two free wills packed together
inside the same cranial vault reminds us that, after con-
sciousness, free will is probably the next most treasured
property of the human brain. Questions and information
relating to the evolution of free will have practical impact
rating right at the top, along with those of consciousness.
As such it probably deserves at least a closing comment.
Some maintain that free will is an evolved, emergent prop-
erty of the brain that appeared between man and the higher
apes, or, depending on whom you read, maybe somewhere
after bacteria perhaps, but before houseflies.

Unlike "mind," "consciousness," and "instinct," "free
will" has made no comeback in behavioral science in recent
years. Most behavioral scientists would refuse to list free
will among our problems outstanding, or at least as an
unanswered problem. (To agree that behavior is unlawful
in this respect might put them out of work as scientists, you
see, and oblige them perhaps to sign up with the astrologers'
union.) Every advance in the science of behavior, whether
it has come from the psychiatrist's couch, from microelec-
trode recording, from brain-splitting, or from the running
of cannibalistic flatworms, seems only to reinforce that old
suspicion that free will is just an illusion. The more we
learn about the brain and behavior, the more deterministic,
lawful, and causal it appears.

In other words, behavioral science tells us that there is
no reason to think that any of us here tonight had any real
choice to be anywhere else, or even to believe in principle
that our presence here was not already "in the cards," so
to speak, five, 10, or 15 years ago. I do not feel comfortable
with this kind of thinking any more than you do, but so far I
have not found any satisfactory way around it. Alternatives
to the rule of causal determinism in behavior proposed so far,
like the inferred unlawfulness in the dance of subatomic
particles, seem decidedly more to be deplored as a solution

than desired.

The above statements are not to say that, in the practice
of behavioral sciences, we must regard the brain as just a
pawn of the physical and chemical forces that play in and
around it. Far from it. To go back to the beginning of the
present lecture, recall that a molecule in many respects is
the master of its inner atoms and electrons. The latter are
hauled and forced about in chemical interactions by the
over-all configurational properties of the whole molecule.
At the same time, if our given molecule is itself part of a
single-celled organism such as Paramecium, it in turn is
obliged, with all its parts and its partners, to follow along
a trail of events in time and space determined largely by
the extrinsic over-all dynamics of Paramecium caudatum.
When it comes to brains, remember that the simpler electric,
atomic, molecular, and cellular forces and laws, though
still present and operating, have been superseded by the
configurational forces of higher-level mechanisms. At the
top, in the human brain, these include the powers of per-
ception, cognition, reason, judgment, and the like, the op-
erational, causal effects and forces of which are equally or
more potent in brain dynamics than are the outclassed inner
chemical forces.

You sense the underlying policy here: "If you can't lick
'em, join 'em," or, as Confucius might say, "If fate inevitable,
relax and enjoy," or, "There may be worse fates than causal
determinism." Maybe, after all, it is better to be embedded
firmly in the causal flow of cosmic forces, as an integral part
thereof, than to be on the loose and out of contact with these
forces, "free floating" as it were and with behavioral possi-
bilities that have no antecedent cause and hence no reason,
nor any reliability when it comes to future plans, predictions,
or promises.

And on this same theme, just one final point : If you were
assigned the task of trying to design and build the perfect

free-will model (let us say the perfect, all-wise, decision-
making machine to top all competitors' decision-making
machines), consider the possibility that your aim might not
be so much to free the machinery from causal contact, as the
opposite, that is, to try to incorporate into your model the
potential value of universal causal contact; in other words,
contact with all related information in proper proportion —
past, present, and future.

It is clear that the human brain has come a long way in
evolution in exactly this direction when you consider the
amount and the kind of causal factors that this multidimen-
sional intracranial vortex draws into itself, scans, and brings
to bear on the process of turning out one of its "preordained
decisions." Potentially included, thanks to memory, are the
events and collected wisdom of most of a human lifetime.
We can also include, given a trip to the library, the accumu-
lated knowledge of all recorded history. And we must add
to all the foregoing, thanks to reason and logic, much of the
future forecast and predictive value extractable from all these
data. Maybe the total falls a bit short of universal causal
contact; maybe it is not even quite up to the kind of thing
that evolution has going for itself over on Galaxy Nine; and
maybe, in spite of all, any decision that comes out is still
predetermined. Nevertheless it still represents a very long
jump in the direction of freedom from the primeval slime
mold, the Jurassic sand dollar, or even the latest 1964-model
orangutan.

LITERATURE CITED

BOGEN, J. E., AND P. J. VOGEL

1962. Cerebral commissurotomy in man. Bull. Los Angeles Neurol. Soc.,
vol. 27, p. 169.

Gazzaniga, M. S., J. E. Bogen, and R. W. Sperry

1962. Some functional effects of sectioning the cerebral commissures in man.

Proc. Natl. Acad. Sci., vol. 48, pt. 2, p. 1765.
1963. Laterality effects in somesthesis following cerebral commissurotomy
in man. Neuropsychologia, vol. 1, p. 209.

Sperrv, R. W.

1950a. Mechanisms of neural maturation. In Stevens, S. S. (ed.), Handbook

of experimental psychology. New York, John Wiley and Sons, p. 236.
1950b. Neural basis of the spontaneous optokinetic response produced by

visual inversion. Jour. Comp. Physchol., vol. 43, p. 482.
1951. Regulative factors in the orderly growth of neural circuits. Growth

Symp., vol. 10, p. 63.
1958. Physiological plasticity and brain circuit theory. In Harlow, H. F., and

C. N. Woolsey (eds.), Biological and biochemical bases of behavior.

Madison, University of Wisconsin Press.

1961. Some developments in brain lesion studies of learning. Fed. Proc,
vol. 20, p. 609.

1962. How a random array of cells can learn to tell whether a straight line
is straight — discussion on. In Foerster, H. von, and G. W. Zopf, Jr.
(eds.), Principles of self-organization. New York, Pergamon Press,
p. 323.

1963. Chemoaffinity in the orderly growth of nerve fiber patterns and con-
nections. Proc. Natl. Acad. Sci., vol. 50, p. 703.

1964. The great cerebral commissure. Sci. Amer., vol. 210, p. 42.

! .4
J35
9 65

AMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

1965

EVOLUTION OF PHYSICAL
CONTROL OF THE BRAIN

JOSfi M. R. DELGADO

f NMURMJJSS2L

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1965

'*/

1869
THE LIBRARY

JAMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

FAMES \k l III K Mt rUR] <>N

I 1 1 I | \ ( ) | l I I ( ) \ O I I 1 1 I HUMAN B K \ I N

l 9 6 5

EVOLUTION OF PHYSICAL
CONTROL OF THE BRAIN

JOSE M. R. DELGADO, M.D.

A ssociate Professor of Physiology

Yale University School of Medicine

New Haven, Connecticut

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1965

JAMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

Frederick Tilney, The Brain in Relation to Behavior; March 15, 1932

C. Judson Herrick, Brains as Instruments of Biological Values; April 6, 1933

D. M. S. Watson, The Story of Fossil Brains from Fish to Man; April 24, 1934
C. U. Ariens Kappers, Structural Principles in the Nervous System; The Develop-
ment of the Forehrain in Animals and Prehistoric Human Races; April 25,
1935

Samuel T. Orton, The Language Area of the Human Brain and Some of its Dis-
orders; May 15, 1936
R. W. Gerard, Dynamic Neural Patterns; April 15, 1937
Franz Weidenreich, The Phylogenetic Development of the Hominid Brain and its

Connection with the Transformation of the Skull; May 5, 1938
G. Kingsley Noble, The Neural Basis of Social Behavior of Vertebrates; May 1 1 ,

1939
John F. Fulton, A Functional Approach to the Evolution of the Primate Brain;

May 27, 1943
James Howard McGregor, The Brain of Primates; May 1 1, 1944
K. S. Lashley, Neural Correlates of Intellect; April 30, 1945
Warren S. McCulloch, Finality and Form in Nervous Activity; May 2, 1946
S. R. Detwiler, Structure-Function Correlations in the Developing Nervous Sys-
tem as Studied by Experimental Methods; May 8, 1947
Tilly Edinger, The Evolution of the Brain; May 20, 1948
Donald O. Hebb, Evolution of Thought and Emotion; April 20, 1949
Ward Campbell Halstead, Brain and Intelligence; April 26, 1950
Harry F. Harlow, The Brain and Learned Behavior; May 10, 1951
Clinton N. Woolsey, Sensorv and Motor Systems of the Cerebral Cortex; May 7,

1952
Alfred S. Romer, Brain Evolution in the Light of Vertebrate History; May 21,

1953
Horace W. Magoun, Regulatory Functions of the Brain Stem; May 5, 1954
Fred A. Mettler, Culture and the Structural Evolution of the Neural System;

April 21, 1955
Pinckney J. Harman, Paleoneurologic, Neoneurologic, and Ontogenetic Aspects

of Brain Phylogeny; April 26, 1956
Davenport Hooker, Evidence of Prenatal Function of the Central Nervous System

1963
Roger W. Sperry, Problems Outstanding in the Evolution of Brain Function;

June 3, 1964
Jose M. R. Delgado, Evolution of Physical Control of the Brain; May 6, 1965

EVOLUTION OF PHYSICAL
CONTROL OF THE BRAIN

INTRODUCTION

I would like to express my gratitude for the privilege of
addressing this distinguished audienee, and also my feeling
of responsibility in following so many illustrious predeces-
sors and in honoring the founder of the James Arthur Lec-
tures on the Evolution of the Human Brain. The topics
covered by earlier speakers in this series have included be-
havioral implications derived from cerebral anatomy and
physiology, neurophysiological problems, comparative
anatomy, embryology, and fossil skulls. In this year's lec-
ture, I would like to project cerebral evolution toward the
future without losing touch with the solid ground of ex-
perimentation.

The human brain has evolved with a functional asymmetry
which may be responsible for some of the conflicts of our
present age. Apparently it has been easier for man to direct
his attention outward to the environment than inward to
deal with the complexity of his own mental structure, and
easier to understand and manipulate Nature than to control
his own behavior. In prehistoric times, and even today in
primitive societies, man was and is at the mercy of the ele-
ments. When disaster struck, and floods, pestilence, or hun-
ger desolated the land, the only possible reactions were
fatalistic resignation, appeal to supernatural powers, or
despair. Modern civilization has progressed so much in the
understanding and domination of the physical world, that
relations between man and Nature have been completely
transformed. Technology is reshaping the face of the earth,

but the greatest change has taken place in the human brain
which is now filled with new formulas, theories, and knowl-
edge, and empowered with a new attitude of confidence
toward natural forces which are no longer the masters but
are becoming the servants of man. The expanding sciences
have directed most of our present intellectual and economic
power toward industry, biology, electronics, atomic energy,
outer space, and similar fields of endeavor, while only a
minor fraction is devoted to inquiry into the roots of men-
tal faculties. This unbalanced interest has an explanation.
When observation and reason were the main tools for the
acquisition of knowledge, philosophical speculation flour-
ished. When the discovery of new methods permitted the
scientific exploration of Nature, the study of subjects beyond
experimental reach was neglected. Certainly, the disciplines
of psychology and psychiatry have greatly expanded in our
century, but a perusal of the literature shows that until one
or two decades ago, the brain was treated as a "black box"
which could be reached only through the senses. Psycho-
logical investigations analyzed correlations between sensory
input and behavioral output, but it was not possible to ex-
plore the processes lying in between which were hidden in
the mystery of brain physiology.

During the last decade we have reached an historical
turning point because of the development of methods which
permit the coordination and synthesis of physical, physio-
logical, pharmacological, and psychological research. As
will be explained in the following pages, science has devel-
oped a new electrical methodology for the study and con-
trol of cerebral functions in animals and humans. Learning,
emotions, drives, memory, consciousness, and other phe-
nomena which in the past belonged only in the realm of
philosophy are now the subjects of neurophysiological ex-
perimentation. In the last few years, the scalpel of the brain

surgeon has modified psychological reactions and a wealth

of wonder drugs has liberated many patients from mental
institutions.

I am not so naive as to think that cerebral research holds
all the answers to mankind's present problems, but I do
believe that an understanding of the biological bases of
social and antisocial behavior and of mental activities,
which for the first time in history can now be explored in
the conscious brain, may be of decisive importance in the
search for intelligent solutions to some of our present anxie-
ties, frustrations, and conflicts. Also, it is essential to intro-
duce a balance into the future development of the human
mind, and I think that we now have the means to investi-
gate and to influence our own intellect.

In support of these ideas, I shall present a brief outline
of the evolution of the physical control of cerebral proc-
esses, followed by several examples of our incipient control
of behavioral mechanisms, and I will end with a discussion
of the principles and implications involved.

HISTORICAL OUTLINE: THEORETICAL
AND METHODOLOGICAL EVOLUTION

Animal Experimentation

For many centures it was accepted that fluids or "animal
spirits" were the cause of muscle contraction (Galen, 130
to ca. 200 a.d.), until the famous controversy between the
schools of Luigi Galvani (1737-1789) and Alessandro
Volta (1745-1827) focused the attention of nineteenth-
century scientists and philosophers on the possible physical
control of some manifestations of life. Contractions pro-
duced in a frog nerve-muscle preparation by touching it
with a bimetallic arc were interpreted by Galvani as proof
of the existence of animal electricity, while Volta believed
that the electrical source was in the contact of two dissimilar

metals. This controversy was resolved when Alexander
von Humboldt (1769-1859) demonstrated that animal
electricity and bimetallic electricity were co-existing phe-
nomena. Leg movements evoked in frogs by the inanimate
force of electricity proved that muscle contraction could be
induced independently of the "principle of life" which had
been considered the essential mover of all biological activi-
ties. The discovery that living organs could be influenced by
instrumental manipulations directed by the will of a human
being brought about a revision of the traditional concepts
of vitalism which were challenged at that time by Emil
DuBois-Reymond (1818-1896) and other scientists. The
romantic mystery of the soul's "animal spirits" which had
dominated biology for almost 2000 years now gave place
to more prosaic chemical and physical laws, and even ner-
vous activity could be investigated experimentally. DuBois-
Reymond not only discovered many basic neurophysiolog-
ical principles, including action current, polarization,
electrogenesis, and propagation of the nerve impulse; he also
provided the technical means for study of the two most
fundamental processes of neural activity by inventing the
galvanometer for the detection of electrical currents and the
induction coil for faradic excitation of nervous tissue. At
that time, the possibility of exciting the spinal cord and brain
stem by other than physiological stimuli was violently de-
bated, and the excitability of the brain was completely
denied. Then Fritsch and Hitzig (1870) performed a beau-
tiful series of experiments, applying galvanic stimulations
to the exposed cerebral cortex of anesthetized dogs. Excita-
tions of the posterior part of the brain failed to evoke
motor effects, but in the anterior region contralateral body
and limb movements were elicited. Weak currents induced
discrete contractions localized to specified muscle groups,
while stronger currents increased the strength and spread

of the evoked responses; it' the intensity was further aug-
mented, generalized convulsions appeared.

The scientific impact of these studies, and also the suc-
cessful clinical localization of speech functions by Broca
( 1824-1880), promoted great interest in cerebral mapping,
based on regional ablation and electrical stimulation studies
attempting to pin precise functional labels to specific ana-
tomical structures. Fortunately, there was much less specu-
lation and much more experimentation in these studies than
in the discredited phrenology, and, in spite of controversial
issues, many of the facts discovered in the last century have
remained important scientific contributions.

One of the main handicaps in these investigations was
the need for opening the skull and exposing the brain.
Operations were usually performed under general anesthesia
which blocked pain perception but also blocked some of
the most important functions of the nervous system. Emo-
tions, consciousness, and intelligence were certainly absent
in heavily sedated animals or in the isolated nerves of the
squid, and for many years scientists directed their atten-
tion to sleeping brains and overlooked the complexity of
awake minds. Textbooks of cerebral physiology were con-
cerned with synapses, pathways, reflexes, posture, and
movement, while mental functions and behavior were con-
sidered to belong to a different discipline.

Some pioneer efforts, however, were directed toward
exploration of the waking brain, and techniques were de-
vised for the introduction of wires through the skull in order
to apply electrical currents to the brains of conscious ani-
mals. In 1898, Ewald had the idea of screwing an ivory
cone into the skull of an anesthetized dog, and the following
day, when operative anesthesia had worn off, electrodes
were inserted into the brain through the ivory piece. A leash
around the animal's neck contained stimulating wires, and

a small dry-cell battery carried by the observer served as
the electrical source. Although the technique and results
were primitive, a way had been found to investigate the
brain in awake animals. The technique of intracerebral
electrodes was dormant for many years until Hess (1932)
developed his own method to explore the hypothalamus
and other cerebral areas in unanesthetized cats. In a series
of brilliant experiments, Hess demonstrated that autonomic
functions, posture, equilibrium, movement, sleep, and even
fear and aggressiveness may be influenced by electrical
stimulation of specific cerebral structures. For the first
time, it was revealed that psychological manifestations like
rage do not depend exclusively on sensory inputs and physi-
ological stimuli, but can be induced by electrical currents
applied directly to the brain. Although these findings did
not produce a significant impact on philosophical thinking,
in retrospect they may be considered as important as the
nineteenth-century demonstration that the contraction of a
frog muscle did not depend on circulating spirits and could
be controlled by physical instrumentation.

For two decades, the methods of Hess attracted only
limited interest among biologists, but in the 1950's, there
was a sudden expansion of the new disciplines of psycho-
surgery, psychosomatic medicine, psychopharmacology,
and physiological psychology, and many investigators real-
ized the great research potential of intracerebral methods
for the study of behavioral-cerebral correlations in awake
animals. With this increased interest, a variety of technical
improvements appeared. Electrodes were no longer intro-
duced free-hand into the brain, but were inserted with
geometric precision with the aid of micromanipulators and
stereotaxic coordinates. Anatomical maps of the depths of
the brain were compiled for rats, cats, dogs, and monkeys.
Aseptic precautions and instrumental refinements permitted

long-term implantation of electrodes, which in some cases
lasted for several years. The sight of experimental animals
with sockets on top of their heads was exceptional in 1950
but had spread to hundreds of laboratories around the
world by 1960. Electrodes were implanted not only in the
usual laboratory animals, but also in other species, includ-
ing crickets, roosters, chimpanzees, dolphins, and brave
bulls.

Experiments were generally performed under some re-
straint. Rats were convenient subjects because of their
behavioral simplicity, and they were not disturbed by a
light coil of wires connecting their terminal head sockets
with the stimulators. In this way, the brain was stimulated
in fully conscious rats while they pressed levers, ran mazes,
and maneuvered with considerable freedom, being limited
only by the length of the leads and the size of the cage. A
similar set-up was also used successfully with cats, provid-
ing they were peaceful and tame. These studies were often
extended for months and were very appropriate for the
investigation of autonomic, somatic, and behavioral effects
evoked by electrical stimulation of the brain, and also for
the analysis of electrical recordings taken during spon-
taneous or induced activities. The combination of intra-
cerebral electrodes with other physiological and psycholog-
ical techniques was very fruitful and showed that animals
can learn to perform instrumental responses to seek or avoid
stimulation of determined cerebral structures. Scientific
exploitation of these techniques continues today with uni-
versal acceptance, as shown by current scientific literature.

The use of electrodes in monkeys presented a greater
challenge because of their destructive skills and restless
curiosity. A heavy protection of the connecting leads was
necessary when the animal was observed on a testing table.
In other cases, the monkey was placed in a special restrain-

ing chair where it could manipulate levers and feed itself
without being able to reach the terminal sockets on its
head. In these situations, conditioning and psychological
testing were successfully performed, but spontaneous be-
havior was naturally curtailed.

The connecting leads trailing behind each animal were
a serious handicap for behavioral studies and were unsuita-
ble for use in chronic stimulations or investigations of group
activities. The obvious solution was to use remote-controlled
instrumentation, with a receiver carried by the animal and
activated by induction or by radio. Several stimulators of
this type have been proposed in the last 30 years (see
bibliography in Delgado, 1963b), but solutions to many
of the technical problems involved were not found until
recently, when the development of transistors and elec-
tronic miniaturization permitted the construction of small,
practical, and reliable cerebral radio stimulators (Delgado,
1963b). After a considerable amount of trial and error, and
in spite of the primates' genius for destroying any equip-
ment within reach, monkey-proofing of instruments was
achieved (figs. 3, 4, 5 ) . The use of radio stimulators allowed
the excitation of cerebral structures in completely free ani-
mals engaged in normal activities within an established
colony and unaware of the scientist's manipulations. In this
way, the role of specific areas of the brain in social relations
was investigated. At the same time, blood pressure, body
temperature, electrical activity of the heart and brain, and
other physiological variables could be recorded by radio
telemetry. In addition, individual and social behavior have
been continuously recorded, day and night, by time-lapse
photography. Radio techniques represented an important
step toward physical control of the brain, providing an
essential tool for behavioral studies, and it may safely be
predicted that within a few years telestimulation will spread

to most brain research institutes. We can also expect that
new developments in micro-electronics, including integrated
circuits and thin film techniques, will facilitate the eonstruc-
ton of multi-channel radio-activated stimulators reduced in
size to a few millimeters. The limits of brain control do not
seem to depend on electronic technology but on the biolog-
ical properties of living neurons.

Among possible physiological handicaps, the presence of
electrodes and repeated applications of electricity could be
disrupting factors for the normality of the nerve cells. Inser-
tion of electrodes into the brain substance is certainly a
traumatic procedure which destroys neural tissue and pro-
duces local hemorrhage, followed by inflammation, foreign
body reaction, and the formation of a glial capsule 0.1-
0.2 mm. thick around the inserted wires. All of this reac-
tive process is limited to a very small area measured in
tenths of millimeters, and there is no evidence of functional
disturbance in the neighboring neurons. Beyond the elec-
trode tract, the brain appears histologically normal and
electrodes seem to be well tolerated, as judged by the
absence of abnormal electrical activity, by the reliability of
effects evoked by electrical stimulation, and by the con-
sistency of thresholds through months of experimentation
(Delgado, 1955b). The longest reported implantation time
of electrodes in the brain has been over four years, in a
rhesus monkey.

From the functional point of view, two aspects should be
considered in implantation experiments. The first is related
to fatigability and the second to lasting functional changes.
Physiological textbooks state that motor effects produced
by electrical stimulation of the cerebral cortex fade away
in a few seconds, and that a rest period of about one minute
is necessary before the cortex recovers its excitability. If
this were true throughout the brain, electricity could not

be effectively used for control of cerebral function. How-
ever, experimentation has shown that the fatigability of
some areas is slow or negligible. In monkeys, the putamen
has been stimulated for more than 30 minutes without
diminution of the elicited postural changes, and the hy-
pothalamus has been excited for days without fatigue of
the evoked pupillary constriction. Red nucleus stimulation
repeated every minute for 14 days has evoked reliable and
consistent sequential responses. Thus, while a few areas of
the brain show quick fatigability, it should be recognized
that many others can be stimulated effectively for minutes
or even days. The evoked effects generally have lasted only
as long as the stimulation, but in some cases enduring after-
effects have been obtained. In the cat, programmed inter-
mittent stimulations of the amygdala for one hour daily
evoked bursts of high-voltage fast activity and other signs
of increased electrical activity, along with changes in spon-
taneous behavior which outlasted stimulation periods for
many hours and occasionally for days. In other studies, ex-
citation of the basolateral nucleus of the cat's amygdala
for only 10 seconds inhibited food intake for minutes, and,
in one case, the inhibitory effect persisted for three days
(Fonberg and Delgado, 1961 ). These findings together with
extensive experimentation by many authors have demon-
strated that intracerebral electrodes are safe and can be
tolerated for years, providing an effective tool for sending
and recording electrical impulses to and from the brain
of unanesthetized animals.

Electrodes in the Human Brain

With the background of animal experimentation, it
was natural that some investigators should contemplate the
implantation of electrodes inside the human brain. Neuro-
surgeons had already proved that the central nervous sys-

teni is not so delicate as most people believe, and during
therapeutic surgery parts of cerebral tissue had been cut,
frozen, cauterized, or ablated with negligible adverse effects
on the patient. Exploratory introduction of needles into
the cerebral ventricles was a well-known and relatively safe
clinical procedure, and, as electrodes are smaller in diam-
eter than these needles, their introduction into the brain
tissue should be even less traumatic. Implantation of elec-
trodes inside the human brain offered the opportunity for
prolonged electrical exploration which could be decisive
for several diagnostic and therapeutic procedures. For ex-
ample, when brain surgery and ablation are contemplated
in patients suffering from epileptic attacks, it is essential
to identify the focal areas of abnormal electrical activity.
Electrodes may remain in place for days or weeks, during
which spontaneous seizures can be recorded and detailed
exploration repeated as many times as necessary. In other
cases, intracerebral electrodes have been used to deliver
intermittent stimulations for periods of days or even months
(Feindel, 1961; Heath, 1954; King, 1961; Sem-Jacobsen
et aL, 1956; Walker and Marshall, 1961). Similar pro-
cedures have also been used in patients with intractable
pain, anxiety neurosis, and involuntary movement. These
therapeutic possibilities should be considered rather tenta-
tive, but accumulated experience has shown that electrodes
are well tolerated by the human brain for periods of at
least one year and a half, and that electrical stimulations
may induce a variety of responses, including changes in
mental functions, as will be explained later. The prospect
of leaving wires inside the thinking brain could seem bar-
baric, uncomfortable, and dangerous, but actually the pa-
tients who have undergone this experience have had no ill
effects, and they have not been concerned about the idea
of being wired or by the existence of leads in their heads.

In some cases, they enjoyed a normal life as out-patients,
returning to the clinic for periodic stimulations. Some of the
women proved the adaptability of the feminine spirit to all
situations by designing pretty hats to conceal their electrical
headgear.

The use of electrodes in the human brain is part of the
present medical orientation toward activation of physiolog-
ical mechanisms by electronic instrumentation, which al-
ready extends to several organs of the body. The clinical
success of electrical driving of cardiac functions in man
has been widely acclaimed. In spite of the delicacy and
continuous mobility of the heart, stainless steel leads have
been sutured to it, and in cases of block in the cardiac
conduction system, artificial electronic pacemakers have
been able to regulate heart rhythm, saving the lives of
many patients. The bladder has been stimulated by im-
planted electrodes to induce urination in patients with
permanent spinal block, and paralyzed limbs have been
activated by programmed stimulators. A method has re-
cently been described for placing leads in the auditory
nerve to circumvent deafness caused by inner ear damage.
Driving malfunctioning organs is simpler than attempting
to direct the awake brain where millions of neurons are
functioning and firing simultaneously for different pur-
poses, but the expected results in this case are even more
interesting. Exploring intracerebral physiology, we are
reaching not only the soma but also for the psyche itself.
Cerebral functions are usually classified in three groups:
autonomic, somatic, and psychic, and in the following
pages I shall discuss present experimental evidence for their
electrical control.

I \BLE OF HISIORK \I EVOLUTION
OK PHYSK \i C'ONFROI OF IHI BRAIN

Findings

Frog muscle contracted when stimu-
lated b\ electricity. Volta. Galvani,
DuBoa-Reymond; 1780, 1800. 1848

Electrical stimulation of the brain in
anesthetized dog evoked localized
body and limb movements. Fritscfa
and Hitzig. 1870

Stimulation of the diencephalon in un-
anesthetized cats evoked well-or-
ganized motor effects and emotional
reactions. Hess. 1932

In single animals, learning, condition-
ing, instrumental responses, pain,
and pleasure have been evoked or
inhibited by electrical stimulation of
the brain in rats. cats, and monkeys.
See bibliography in Sheer. 1961

In colonies of cats and monke\s. ag-
gression, dominance, mounting, and
other social interactions have been
evoked, modified, or inhibited by
radio stimulation of specific cerebral
areas. Delgado. 1955a. 1964

In patients, brain stimulations during
surgical interventions or with elec-
trodes implanted for days or months
have blocked the thinking process,
inhibited speech and movement, or
in other cases have evoked pleasure,
laughter, friendliness, verbal output,
hostility, fear, hallucinations, and
memories. See bibliography in
Ramev and O'Dohertv. 1960

Imim k \i IONS

"Vital spirits" are not essential for bio-
logical activities. Electrical stimuli
under man's control can initiate and
modify vital processes

The brain is excitable. Electrical stim-
uli of the cerebral cortex can pro-
duce movements

Motor and emotional manifestations
may be evoked by electrical stimula-
tion of the brain in awake animals

Psychological phenomena may be con-
trolled by electrical stimulation of
specific areas of the brain

Social behavior may be controlled by
radio stimulation of specific areas of
the brain

Human mental functions may be in-
fluenced by electrical stimulation of
specific areas of the brain

Summary: Autonomic and somatic functions, individual and social behavior,
emotional and mental reactions may be evoked, maintained, modified, or
inhibited, both in animals and in man. by electrical stimulation of specific
cerebral structures. Physical control of many brain functions is a demon-
strated fact, but the possibilities and limits of this control are still little known.

ELECTRICAL CONTROL OF AUTONOMIC FUNCTIONS

Several areas of the brain play important roles in the
regulation of visceral activity, and extensive studies have
shown that electrical stimulation of the hypothalamus and
other cerebral structures can influence vasomotility, blood
pressure, heart rate, respiration, thermal regulation, gastric
secretion, food intake, and many other functions of the
autonomic system. To illustrate the artificial regulation of
autonomic reactions by electrical means, I shall discuss
pupillary motility because its mechanisms are relatively
simple and easy to control.

The areas that participate in the regulation of pupil size
are represented on the surface and in the depth of the brain.
Cortical zones which have inhibitory effects upon respira-
tion and upon spontaneous movements also produce pupil-
lary dilatation (mydriasis). In cats, dogs, and monkeys,
these areas are situated around the sylvian fissure, orbital
cortex, temporal tip, cingulate gyrus, insula, rhinal fissure,
and hippocampal gyrus. In the depth of the brain, pupillary
dilatation may be evoked by stimulation of the basal telen-
cephalon, hypothalamus, septum, midline group of thalamic
nuclei, subthalamus, and a large part of the midbrain
(Hodes and Magoun, 1942; Kaada, 1951; Showers and
Crosby, 1958). Pupillary constriction (miosis) has a more
limited representation, localized mainly around the genu
of the corpus callosum (Hodes and Magoun, 1942; Kaada,
1951), thalamus, and hypothalamus (Hess, 1954). Accord-
ing to the region stimulated, pupillary responses will be
unilateral or bilateral; if bilateral, each eye may respond
synergically or antagonically. Most classical studies were
performed under anesthesia and with the brain exposed,
but recent investigations have been carried out with the
use of awake animals equipped with intracerebral electrodes.

In monkeys (Delgado, 1959), electrical stimulation of

Fig. 1. The diameter of the pupil may be electrically controlled as if it was
the diaphragm of a photographic camera. The pictures show normal eyes in a
monkey and the dilatation and constriction of the right pupil evoked by stimula-
tion of the hypothalamus. Some of these effects are indefatigable and persist for
days as long as stimulation is applied.

the inferior part of the lateral hypothalamus produced
marked ipsilateral miosis, while stimulation of another
point situated 6 mm. higher in the same tract evoked ipsi-
lateral mydriasis (fig. 1). The magnitude of the effect was
proportional to the electrical intensity employed. Stimula-
tion of the inferior point with 0.8 milliampere (mA) pro-
duced slight pupillary constriction which increased progres-
sively as the intensity was augmented to 1.5 mA. At this
moment, miosis was maximum, and further increase in
stimulation did not modify the effect. If the hypothalamic
stimulation was slowly decreased in strength, the ipsilateral
pupil gradually returned to its normal size. In these experi-
ments, pupil diameter could be controlled precisely like
the diaphragm of a camera, by turning the stimulator dials
to the left or right. A similar dose-response relation was
seen in the higher hypothalamic point where stimulation
produced mydriasis. Implantation of electrodes in points
with antagonistic pupillary effect made it possible to intro-
duce an artificial conflict by stimulating both areas simultane-
ously with separate instruments. Results showed that a
dynamic equilibrium could be established at different levels
of simultaneous antagonistic excitation. With 1.6 foot-
candle units of illumination in the laboratory, the initial
pupillary diameter of 4 mm. was maintained when the
hypothalamic points were stimulated together at similarly
increasing intensities up to 4 mA. At any level in this
dynamic equilibrium, the pupil constricted if intensity was
increased in the inferior or decreased in the higher point.
The reverse was also true, and the pupil dilated if stimula-
tion decreased in the inferior or increased in the superior
hypothalamic point. To some extent, the effect of excitation
of the inferior miotic point could be substituted for a light
shone in the eye, illustrating the possibility of algebraic
summation of physiological, sensory, and electrical stimuli

within the brain. These experiments demonstrated that a
regulation of an autonomic- function like pupillary size can
be effectively maintained by direct stimulation of cerebral
structures.

For how long would this regulation be effective? Would
the brain fatigue? To answer these questions, long-term
experiments were designed. Under continuous hypothalamic
excitation, mydriasis lasted for about 30-40 minutes, after
which stimulation was ineffective and the pupil gradually
returned to its original size, indicating a slow fatigability of
the effect. In contrast, pupillary miosis was maintained in
several monkeys for as long as stimulation was applied.
Each Animal was studied while free in a cage and equipped
with a portable stimulator connected by subcutaneous
leads to the inferior hypothalamic point. Under continuous
24-hour stimulation, the size of the ipsilateral pupil was
maintained at less than 1 mm. in diameter, while the other
pupil measured a normal 4 mm. As soon as the stimulation
was discontinued, a rebound effect appeared and the ipsi-
lateral pupil dilated to about 6 mm. for several hours, and
then slowly returned to its normal size. In one monkey, the
stimulation was applied for as long as three days, during
which pupil constriction was continuous; with cessation of
stimulation, a rebound effect appeared which lasted for
two days.

In other experiments, when the intensity of hypothalamic
stimulation was adjusted to produce only a 20-30 per cent
reduction in pupillary size, the reactivity of both pupils to
light was preserved, although the stimulated pupil was al-
ways smaller than the control. These results demonstrated
that a lasting functional "bias" can be introduced in auto-
nomic reactions by the artificial means of electrical stimu-
lation of the brain. The physiological equilibrium was elec-
trically modified, preserving the responses but changing the

level of functional adjustment. These results are compar-
able to the modifications in autonomic reactivity (tuning)
induced by injection of sympathetic or parasympathetic
agents (Gellhorn, 1957).

In summary, autonomic functions can be controlled by
electrical stimulation of the brain. As an example it has
been shown that constriction of the pupil evoked by cere-
bral stimulation is reliable, precise, does not fatigue, can
interplay with physiological stimuli, and may provide a
functional "bias" to modify the level of physiological re-
sponses.

MOTOR PERFORMANCE UNDER
ELECTRONIC COMMAND

The significant nineteenth-century discovery of central
nervous system excitability was based on the fact that elec-
trical stimulation of the cerebral cortex produced observable
motor responses. Since that discovery, many investigations
have been devoted to the analysis of motor representation
in different areas of the brain. The evoked effects were
usually described as stereotyped tetanic contractions, pro-
ducing clumsy movements of the body and extremities and
lacking the precision and coordination of spontaneous ac-
tivities. These results were obtained under anesthesia, but
it was assumed that because of the complexity of the
mechanisms involved, artificial stimulation could never
induce, even in awake animals, responses as skillful and well
organized as voluntary movements. In spite of this assump-
tion, when stimulation was applied through intracerebral
electrodes to completely unrestrained animals, it was evi-
dent that motor performance under electronic command
could be as complex and precise as spontaneous behavior.
Before discussing the reasons for success in the electric
driving of behavior, I will describe examples of simple
motor responses, complex behavior, and social interaction.

- -

— u
u c
u

I) 3

i i

o "a

a 5

c —

Simple Motor Responses
In the cat, electrical stimulation of the right sulcus cru-
ciatus, in the anterior part of the brain, produced flexion
of the left hind leg (fig. 2) with an amplitude of movement
proportional to stimulation intensity, provided the experi-
mental situation was constant. For example, in a cat stand-
ing on all fours, a five-second stimulation of 1.2 mA (mono-
polar, cathodal, square waves, 0.5 millisecond of pulse
duration, 100 cycles per second) evoked a leg flexion barely
off the ground. When the intensity was increased up to
1.5 mA, the hind leg rose about 4 centimeters, and when
1 .8 mA were applied, the flexion of the leg was complete.
The evoked movement usually began slowly, developed
smoothly, reached its peak in about two seconds, and lasted
until the end of the stimulation. This motor performance
could be repeated as many times as desired, and it was
accompanied by a postural adjustment of the whole body
which included a lowering of the head, raising of the pelvis,
and a slight weight shift to the left in order to maintain
equilibrium on only three legs. The electrical stimulation
did not produce any emotional disturbance, and the cat was
as alert and friendly as usual, rubbing its head against the
experimenter, seeking to be petted, and purring. However,
if we tried to prevent the evoked effect by holding the left
hind leg with our hands, the cat stopped purring, struggled
to get free, and shook its leg. Apparently the evoked motility
was not unpleasant, but attempts to prevent it were dis-
turbing for the animal. The artificial driving of motor ac-
tivities was accepted in such a natural way by the animal
that often there was spontaneous initiative to cooperate
with the electrical command. For example, during a
moment of precarious balance when all paws were close
together, stimulation produced first a postural adjustment,
and the cat spread its forelegs to achieve equilibrium by

shifting its body weight to the right, and only after this
delay did the left hind leg begin to flex. It was evident that
the animal was not in a hurry and was taking its time to
prepare its position for the indueed movement. Preliminary
adjustments were not seen if the cat's posture was already
adequate for the required motor performance. In other
cases, when the animal was lying down with its hind legs
already flexed, the stimulation effect was greatly diminished
and consisted mainly of increased muscular tension.

In cases of conflict between the free movements of the
animal and those elicited by the experimenter, the final
result depended on the relative strength of opposing sig-
nals. Stimulations of the cruciate sulcus at threshold level
of 1.2 mA, which produced a small leg flexion, were in-
effective if applied while the cat was walking. To test
stronger conflicts, the cat was enticed into jumping off a
table to reach food placed on the floor, and, while it was
in the air, the cruciate sulcus was electrically stimulated.
In this situation, intensities of up to 1.5 mA, which usually
evoked a clear motor response, were completely ineffective;
physiological activity seemed to override the artificial ex-
citation and the cat landed with perfectly coordinated move-
ments. If the intensity was increased to 2 mA, stimulation
effects were prepotent over voluntary activities; leg flexion
started during the jump, coordination was disrupted, and
the cat landed badly.

A variety of motor effects have been evoked in different
species, including cat, dog, bull, and monkey. The animals
could be induced to move the legs, raise or lower the body,
open or close the mouth, walk or lie still, turn around, and
perform a variety of responses with predictable reliability,
as if they were electronic toys under human control (see
figs. 1-6). Behavior elicited by electrical stimulation was
not always comparable to spontaneous activity. In a few

experiments, movements beyond the animal's voluntary
control were observed, such as the clockwise rotation of
the eye. In other cases, abnormal responses, disorganized
contractions, and loss of equilibrium have also been in-
duced, depending on the cerebral area and parameters of
stimulation.

Complex Behavior

Normal activities in animals are not confined to simple
motor responses such as hind-leg flexion but include a suc-
cession of different acts such as body displacement and
social interaction. In order to study these complex activi-
ties, which require a situation as free and normal as pos-
sible, our experimental design included ( 1 ) the establish-
ment of a colony with four to six monkeys, (2) the con-
tinuous recording of spontaneous and evoked behavior by
time-lapse photography, in order to qualify and quantify
individual and social actions, and (3) stimulation of the
animals by remote control. The behavior of a group of
monkeys is an entertaining spectacle, and a few minutes'
observation gives the impression that their playing, groom-
ing, chasing, and comic activities are rather unpredictable.
Long-term studies, however, have shown that individual
and social behavior is predictable within a known range of
variability. The study of group behavior is possible pre-
cisely because of the recurrence of patterns that can be
identified. Every day the monkeys will eat, play, groom,
pick, sit, and perform a series of acts which can be analyzed
and quantified (Delgado, 1962). After the individual pro-
files of behavior are established, the responses evoked by
electrical stimulation of the brain may be precisely evalu-
ated.

A typical example of complex behavior was observed
in a monkey named Ludi while she was forming part of a

Fig. 3. Yawning evoked in the monkey by radio stimulation of the pars magno
cellularis of the red nucleus. Observe the spontaneous qualities of the evoked
effect and also the fact that when the monkey is asleep the response diminishes.

colony with two other females and two males. Ludi was an
aggressive female who dominated the whole group and
exercised the usual prerogatives of being the chief, enjoying
greater territoriality and more food, and moving freely
around the colony. After different areas of the brain had
been studied under restraint, the radio stimulator was
strapped to Ludi, and excitations of the rostral part of the
red nucleus were started, with the monkey free in her
colony. Stimulation produced the following complex se-
quence of responses (fig. 4): (1) immediate interruption
of spontaneous activities, (2) change in facial expression,
(3) head turning to the right, (4) standing on two feet,
(5) circling to the right, (6) walking on two feet with
perfect preservation of equilibrium by balancing the arms,
touching the walls of the cage, or grasping the swings, (7)
climbing a pole on the back wall of the cage, (8) descend-
ing to the floor, (9) low tone vocalization, (10) threatening
attitude directed toward subordinate monkeys, (11) chang-
ing of attitude and peacefully approaching some other
members of the colony, and (12) resumption of the activity
interrupted by the stimulation. The whole sequence was
repeated again and again, as many times as the red nucleus
was stimulated. Responses 1 to 8 developed during the five
seconds of stimulation and were followed, as aftereffects, by
responses 9 to 12 which lasted from five to 10 seconds. The
excitations were repeated every minute for one hour, and
results were highly consistent on different days. The re-
sponses resembled spontaneous activities, were well organ-
ized, and always maintained the described sequence. Climb-
ing followed but never preceded turning of the body; vocal-
ization followed but never preceded walking on two feet; the
general pattern was similar in different stimulations, but the
details of motor performance varied and were adjusted to
existing circumstances. For example, if the stimulation

E 5
< y

surprised the animal with one arm around the vertical pole
in the cage, the first part of the evoked response was to
withdraw the arm in order to make the turn possible. While
walking on two feet, the monkey was well oriented and was
able to avoid obstacles in its path and to react according to
the social situation. In some experiments, three monkeys
in the colony were simultaneously radio-stimulated in the
red nucleus, and all three performed the full behavioral
sequence without interfering with one another. Changes in
the experimental situation could modify the evoked re-
sponse, as shown in the case of external threat to the colony.
Waving the catching net or a pair of leather gloves on one
side of the home cage induced a precipitous escape of all
monkeys to the other side. Red-nucleus stimulation applied
at this moment was ineffective and did not interfere with
the escape of the animals. In other experiments, after being
deprived of food for 24 hours, the animals were offered
bananas and oranges which they grabbed and ate voraci-
ously. During this time, Ludi's response to radio stimulation
of the red nucleus was completely absent or was reduced to
only a short turn. In one long experiment, excitation of the
red nucleus was repeated every minute, day and night, for
two weeks, with a total of more than 20,000 stimulations.
The remarkable reliability of responses was demonstrated
throughout the whole period, with the following significant
exception. During the day, monkeys take several naps, and
during the night they have a long period of sleep which is
interrupted by several periods of general activity. Time-
lapse recordings showed that, as the stimulated monkey was
falling asleep, the evoked responses progressively dimin-
ished until only a small head movement remained. As soon
as the stimulated animal awoke, the responses reappeared
with all of their complexity. This finding indicates that the
effects evoked by cerebral stimulation are not inflexible and

rigid, but may adapt to changes in the physiological situa-
tion. Examples of other patterns of sequential behavior
have been evoked by excitation of several diencephalic and
mesencephalic structures (Delgado, 1963a, 1964a, 1964b),
showing that sequential activities are anatomically repre-
sented in several parts of the central nervous system.

SOCIAL INTERACTION

The social interaction of animals requires continuous
mutual adaptation, and activities depend on a variety of
factors, including sensory inputs, problem-solving capacity,
emotional background, previous experience, conditioning,
drives, instincts, and intelligent integration of all these
processes. In spite of the extraordinary complexity of these
supporting mechanisms, there is experimental evidence that
electrical stimulation of specific areas of the brain may
influence social interaction such as contactual relations,
hierarchical situations, submissive manifestations, sexual
activity, aggressive behavior, and social fear. By definition,
this type of research requires at least two animals which can
interact with each other, but the study of groups is naturally
preferable.

In 1928 Hess demonstrated that during electrical stimu-
lation of the periventricular gray matter, cats responded
as if threatened by a dog, with dilatation of the pupils,
flattening of the ears, piloerection, growling, and well-
directed blows with unsheathed claws. Similar offensive-
defensive reactions have been described by several authors
(see bibliography in Delgado, 1964a), but it was debatable
whether the apparently enraged animal was aware of its
own behavior and whether the evoked reactions were pur-
posefully oriented; in other words, if the observed phe-
nomena were true or false rage. Today it is known that
both types of rage may be elicited, depending on the loca-

tion of the stimulated points, and we have conclusive evi-
dence that, in cats and monkeys, well-organized behavior
may be evoked by stimulation of the amygdala, postero-
ventral nucleus of the thalamus, fimbria of the fornix, tectal
area, central gray, and other cerebral structures. The fact
that one animal can be electrically driven to fight against
another has been established (Delgado, 1955a). In this
experiment, stimulation of the tectal area in a male cat
evoked the well-known pattern of offensive-defensive reac-
tions. When this animal was placed on a testing stage in
the company of a larger cat, they enjoyed friendly relations,
lying close to each other and purring happily until the
smaller cat was stimulated in the tectal area. At this mo-
ment, it started growling, unsheathed its claws, and launched
a fierce attack against the larger animal which flattened its
ears, withdrew a few steps, and retaliated with powerful
blows. The fight continued as long as the stimulation was
applied. The effect could be repeated, and the stimulated
cat always took the initiative in spite of the fact that it was
smaller and was always overpowered in the battle. After
several stimulations, a state of mistrust was created be-
tween the two animals, and they watched each other with
hostility.

Similar experiments were repeated later in a colony
formed by six cats. When one of them was radio-stimulated
in the tectal area, it started prowling around looking for
fights with the other subordinate animals, but avoiding
one of them which was the most powerful of the group. It
was evident that brain stimulation had created a state of
increased aggressiveness, but it was also clear that the cat
directed its hostility intelligently, choosing the enemy and
the moment of attack, changing tactics and adapting its
motions to the motor reaction of the attacked animal. In
this case, brain stimulation seemed to determine the affec-

live state of hostility, but the behavioral performance

seemed dependent on the individuality of the stimulated
animal, including its learned skills and previous experi-
ences. Stimulation that increased aggressiveness was usually
tested for only five to 10 seconds, but, as it was important
to determine the fatigability of the effect, a longer experi-
ment was performed by reducing the intensity to a level
which did not evoke overt rage. The experimental subject
was an affectionate cat which usually sought petting and
purred while it was held in the experimenter's arms. When
it was introduced into the colony with five other cats, a low-
intensity radio stimulation of the amygdala was applied
continuously for two hours during which the animal's be-
havior was affected. It withdrew to a corner of the cage and
sat there motionless, uttering barely audible growls from
time to time. If any other cat approached, the stimulated
animal started hissing and threatening, and, if the experi-
menter tried to pet him, the growls increased in intensity
and the animal often spat and hissed. This hostile attitude
disappeared as soon as the stimulation was over, and the
cat became as friendly as before. These experiments demon-
strated that brain stimulation could modify animals' reac-
tions toward normal sensory perceptions by a modulating
of the quality of the responses. The effect was similar to the
modifications of spontaneous behavior observed in normal
emotional states.

Monkeys offer better opportunities than cats for the
study of social interaction because of their more numerous
and skillful spontaneous activities. It is well known that
these animals form autocratic societies, where one estab-
lishes himself as boss of the group, claiming a large amount
of the living quarters as his territory, feeding first, and
being avoided by the others, which usually express their
submissiveness by typical actions such as grimacing,

crouching, and presenting. In several of our monkey colo-
nies, we demonstrated that radio stimulation of the postero-
ventral nucleus of the thalamus and central gray increased
the aggressiveness of the stimulated animal and affected
the social hierarchy. Stimulation of the boss monkey in-
duced well-directed attacks against the other members of
the group, which were chased around and occasionally
bitten, but it was evident that the orientation of the evoked
response was influenced by previous experiences. During
stimulation, the boss usually attacked and chased the male
monkeys which represented a challenge to his authority,
but he did not threaten the female who was his favorite
partner. These results confirmed the finding in cat colonies
that aggressiveness induced by cerebral stimulations was
not blind and automatic, but selective and intelligently
directed.

Rhesus monkeys are destructive and dangerous crea-
tures which do not hesitate to bite anything within reach,
including leads, instrumentation, and occasionally the ex-
perimenter's hands. Would it be possible to tame these
ferocious animals by means of electrical stimulation? To
investigate this question, a monkey was strapped to a
chair where it made faces and threatened the investigator
until the rostral part of the caudate nucleus was electrically
stimulated. At this moment, the monkey lost its aggressive
expression and did not try to grab or bite the experimenter,
who could safely put a finger into its mouth! As soon as
stimulation was discontinued, the monkey was as aggres-
sive as before. Later, similar experiments were repeated
with the monkeys free inside the colony, and it was evident
that their autocratic social structure could be manipulated
by radio stimulation. In one case in which the boss monkey
was excited in the caudate nucleus with 1.5 mA for five
seconds every minute, after several minutes the other mon-

keys started to circulate more freely around the cage, often
in proximity to the boss, and from time to time they
crowded him without fear. The intermittent stimulation
continued for one hour, and during this time the territo-
riality of the boss dropped to zero, his walking time was
diminished, and he performed no aggressive acts against
the other members of the colony. About 12 minutes after
the stimulation hour ended, the boss had reasserted his
authority, and his territoriality seemed to be as well estab-
lished as during the control period. In other experiments,
monkeys instead of investigators controlled the activation
of radio stimulation. In this situation, subordinate animals
learned to press a lever in the cage which triggered stimula-
tion of the boss monkey in the caudate nucleus, inhibiting
his aggressive behavior (fig. 5; Delgado, 1963c). Inhibitory
effects have been demonstrated in several species including
brave bulls, as shown in figure 6 (Delgado, et at., 1964).

A different type of effect was demonstrated in another
monkey colony. Radio stimulation of the nucleus medialis
dorsalis of the thalamus in a female monkey produced a
sequential pattern of behavior characterized by a movement
of the head, walking on all fours, jumping to the back wall
of the cage for two or three seconds, jumping down to the
floor, and walking back to the starting point. At this mo-
ment, she was approached by the boss of the colony and
she stood on all fours, raised her tail and was grasped and
mounted by the boss in a manner indistinguishable from
spontaneous mounting. The entire behavioral sequence was
repeated once every minute following each stimulation, and
a total of 8 1 mountings was recorded in a 90-minute period,
while no other mountings were recorded on the same day.
As is natural in social interaction, the evoked responses
affected not only the animal with cerebral electrodes, but
also other members of the colony.

il" J ft. ~H I

..J j | ^ J .:..—-] "•

"---'of J r '"■->

^, ^**

^•*ffi

T8fc '

Fig. 6. A bull in full charge may be suddenly stopped by radio stimulation of
the anterior part of the thalamus.

ANALYSIS OF EVOKED MOTOR BEHAVIOR

The experimental evidence presented in the previous
pages clearly demonstrates that electrical stimulation of
the brain can induce predictable behavioral performance
similar to spontaneous activities. Understanding the sig-
nificance of these findings requires analysis of the physio-
logical mechanisms involved in voluntary movements. A
simple act such as leg flexion requires the precise and pro-
gressive contraction of several muscles in which the strength,
speed, and amplitude of activation of many motor units are
determined by the processing of messages coming from
joints and muscle spindles integrated with another vast
amount of information circulating through the central ner-
vous system. The complexity of neuronal events is even
greater during performance of sequential responses, in
which timing and motor correlations must be adjusted to
the purpose of the movement and adapted to changes in
the environment. Mechanisms responsible for the physio-
logical excitation of spontaneous motility must be highly
sophisticated. In contrast, electrical stimulation of the brain
is very simple and depends on primitive techniques that
apply a train of pulses without modulation, without code,
without specific meaning, and without feedback to a group
of neurons which by chance are situated within an arti-
ficially created field. In view of the complexity of neuronal
integrations, it is not surprising that a few authors have
downgraded the significance of stimulation effects. How
can we explain the contradiction between the crudeness of
these excitations and the refinement of the responses that
they can elicit?

When considering whether a simple electrical stimulus
could be the cause of the many events of a behavioral re-
sponse, we could ask whether a finger pushing a button to
launch a man into orbit is responsible for the complicated

machinery or for the sequence of operations. Evidently the
finger, like a simple stimulus, is only the trigger of a pro-
grammed series of events, and consequently electrical
charges applied to the brain cannot be accepted as the
direct cause of leg flexion or aggression. The effect of elec-
tricity is simply to depolarize some neural membranes and
to initiate a chain reaction. We must remember that even
at the neuronal level, electrical excitation is not responsible
for the many biochemical, enzymatic, thermal, and elec-
trical processes which accompany the evoked action poten-
tials. Evoked effects, like other chain reactions, depend
more on the functional properties of the activated struc-
tures than on the starter. If electrical stimulation is con-
sidered as a non-specific trigger, our discussion must be
focused on what is triggered. Why do movements start, de-
velop, and end? Which motor mechanisms are involved
within the brain? These basic neurophysiological questions
are very difficult to answer because of our limited knowl-
edge, but at least we now have some new tools to initiate
their study, and experimental hypotheses to guide future
research.

A tentative explanation of some of the mechanisms in-
volved in motor activities has been proposed in the theory
of fragmental representation of behavior (Delgado, 1964a)
which postulates that behavior is organized as fragments
which have anatomical and functional reality within the
brain, where they can be the subject of experimental analy-
sis. The different fragments may be combined in different
sequences like the notes of a melody, resulting in a suc-
cession of motor acts which constitute specific behavioral
categories such as licking, climbing, or walking. The theory
may perhaps be clarified with one example. If I wish to take
a cookie from the table, this wish may be considered as a
force called the starter because it will determine the initia-

tion of a series of motor acts. The starter includes drives,
motivations, emotional perceptions, memories, and other
processes. To take the cookie it is necessary to organize a
motor plan, a mechanical strategy, and to decide among
several motor choices, because the cookie may be taken
with the left or right hand, directly with the mouth, or even
by using the feet if one has simian skills. Choice, strategies,
motor planning, and adjustments depend on a set of cerebral
structures, the organizer, which is different from the set
employed by the starter, because the desire for cookies may
exist in hungry people or in completely paralyzed patients,
and the hands can move and reach the table for many dif-
ferent reasons even if there are no cookies. Finally, the
actual contraction of muscles for the performance of the
selected movement to reach the cookie — for example, using
the right hand — depends on a cerebral set, the performer,
different from the previous two, because motor represen-
tation of hands, mouth, and feet is situated in different
areas of the brain, and the choice of muscle group to be
activated is under the supervision of a given organizer.
Naturally, there is a close correlation among these three basic
mechanisms, and also between them and other cerebral
functions. The concept of a brain center as a visible ana-
tomical locus is unacceptable in modern physiology, but
the participation of a constellation of neuronal groups (a
functional set) in a specific act is more in agreement with
our present knowledge. The functional set may be formed
by the neurons of nuclei far from one another: for instance,
in the cerebellum, motor cortex, pallidum, thalamus, and
red nucleus, forming a circuit in close mutual dependence,
and responsible for a determined act such as picking up a
cookie with the right hand.

If we accept the existence of anatomical representation
of the three functional sets: starter, organizer, and per-

former, it is logical that they can be activated by different
types of triggers, and that the evoked results will be related
to the previous experiences linked to the set. The same set,
evoking a similar behavioral response, may be activated by
physiological stimuli, such as sensory perceptions and idea-
tions, or by artificial stimuli, such as electrical impulses.
Depending on the location of contacts, when we stimulate
the brain through implanted electrodes we can activate the
starter, the organizer, or the performer of different behav-
ioral reactions, so that natural and artificial stimuli may
interplay with one another, as has been experimentally
demonstrated.

These theoretical considerations may facilitate the un-
derstanding of so-called willful, free, or spontaneous ac-
tivity. Obviously, the will is not responsible for the chem-
istry of muscle contraction, for the electrical processes of
neural transmission, or even for the intimate organization
of movements; these phenomena depend on spindle dis-
charges, cerebellar activation, synaptic junctions, reciprocal
inhibitions, and other subconscious mechanisms. Voluntary
activity is initiated by a physiological trigger which acti-
vates a chain of preformed mechanisms which exist inde-
pendently inside the brain. The uniqueness of voluntary
behavior lies in its wealth of starters, each one of which
depends on a vast and unknown integration of past experi-
ences and present receptions. However, the organizers and
performers are probably activated in a similar manner by
the will and by electrical means, providing the possibility
of investigating experimentally some of the basic mecha-
nisms of spontaneous behavior.

One limitation of electrical activation of behavior is the
anatomical variability of the brain. Just as there are ex-
ternal physical differences between individuals, there are
variations in the shape and size of our cerebral structures

which make it impossible to place an electrical contact in
exactly the same location in different subjects. Another
important limitation is functional variability. The organi-
zation of brain physiology depends to a great extent on
individual experience which determines the establishment
of many temporary or permanent associations among
neuronal fields. For example, the sound of a bell is neutral
for a naive animal, but will induce secretion of saliva if it
has previously been paired with food, and stimulation of
the auditory cortex should increase salivary secretion only
in the conditioned animal. Anatomical and functional
variabilities are the bases for the differences in individual
personalities. When we stimulate the motor cortex, we can
predict the appearance of a movement but not the details
of its performance, indicating that the effects elicited by
electrical stimulation of the brain have a statistical but not
an individual determination.

ELECTRICAL DRIVING OF
MENTAL FUNCTIONS IN MAN

Elemental psychic phenomena such as hunger and fear
can be analyzed in both animal and man, but processes like
ideation and imagery that are expressed verbally can be
studied only in human beings. The most extensive informa-
tion on this subject has been obtained by Penfield and his
group (see, for instance, Penfield and Jasper, 1954) dur-
ing surgical operations for epilepsy, tumors, or other ill-
nesses. In these procedures, the brain was exposed under
local anesthesia and stimulated electrically under direct
visual control. More recently, as explained in a previous
section, electrodes have been implanted in the brain for
days or weeks, permitting repeated studies in a relaxed
atmosphere, with the patient in bed or sitting comfortably
in a chair. From Penfield's publications and from implanted-

electrode studies, a considerable amount of information has
demonstrated that brain stimulation may induce anxiety,
fear, hostility, pleasure, feelings of loneliness, distortion of
sensory perception, recollection of the past, hallucinations,
and other psychic manifestations. From all this material, I
shall select several representative examples dealing mainly
with ideation, which is perhaps the most interesting and
least understood of the mental processes.

Speech Increase

Patient A. F. was an 11 -year old boy committed to an
institution because of his uncontrollable epileptic seizures
and destructive behavior (see Higgins et ai, 1956). Since
his response to drugs and treatment was unsatisfactory,
brain surgery was decided upon. To direct the operation,
four electrode assemblies were implanted in the temporal
lobes for six days. During this time, intracerebral activity
was recorded, and several spontaneous seizures were regis-
tered. Exploration of the patient included several tape-
recorded interviews of from one and a half to two hours,
behavioral observations, and 69 intracerebral stimulations.
Study of the collected data indicated the existence of a
focus of abnormality in the left temporal tip, and this area
was successfully removed. Recovery from surgery was un-
eventful, and in a few weeks the boy was able to enjoy a
normal life and return to school. Five years later he was
still seizure-free.

In our investigations, the conversations between patient
and therapist were tape-recorded while the spontaneous
electrical activity of the brain was also being registered,
and programmed stimulations were applied to different
cerebral points. The general procedure was explained to
the patient, but, to avoid possible psychological influences,
he was not informed of the exact moment of the stimula-

tions. To establish behavioral and electrical correlations,
the recorded interviews were transcribed, divided into peri-
ods of two minutes, and analyzed by two independent in-
vestigators who counted the number of words and identified
and quantified the verbal expressions according to 39 dif-
ferent categories. Table 1 shows the stimulation effects on
verbal production. During this interview, the patient was
quiet and spoke only four to 17 words every two minutes.
Whenever point RP 1-2 was stimulated, the patient's atti-
tude changed; he became more animated, and his verbal
output increased sharply to a mean of 88 words per two-
minute period.

TABLE 1

(From Higgins. Mahl, Delgado, and Hamlin. 1956)

Stimulations RP 1-2 (N-7) r-Test

Time interval 2'Postim. 2'Prestim. P-Value

All Others /-Test

Stimulations (7V-7) P-Value
2'Postim. 2'Prestim.

Mean % friendly remarks 6 53 0.02 17 10

Mean N words by patient 17 88 <0.01 4 9 0.15

Mean N words by Int. 43 46 — ' 16 30 >0.30

"Insignificant by inspection.

These effects were repeated seven times, and in each
stimulation the patient appeared to be especially optimistic,
emphasizing the pleasant side of sensory perceptions and
the happy aspects of his memories and ideas, with many
of his comments affectionately directed and personally re-
lated to the therapist. Verbal expression was spontaneous
in character, his usual personal style and phraseology were
preserved, and conversational topics were related to the
experimental situation without a preferred theme. Table 1
shows that the evoked increase of words and of friendly
remarks were highly significant, as evaluated by the /-test,
and also that the effect was specific because it was not pro-
duced by stimulation of other cerebral points.

Si \ial Ideation

In three different patients, thoughts and expressions with
sexual eontent were indueed by electrical stimulation of
the temporal lobe. The first case, S. S., was an intelligent
and attractive woman. 32 years old, who had suffered from
uncontrollable epileptic attacks for several years. During
the interviews she was usually reserved, but the first time
that point A in the second temporal convulsion was excited
with 6 volts, she became visibily affected, holding the hands
of the therapist to express her fondness for him and to
thank him for all his efforts. Several minutes later, after
another stimulation of the same point, she started to say
how much she would like to be cured so that she might
marry, and other stimulations of point A were also followed
by flirtatious conversation. The provocative play and ideas
expressed under stimulation of point A did not appear fol-
lowing stimulation of other cerebral points and contrasted
with this woman's usually reserved spontaneous behavior.

The second patient, V. P., was a woman 36 years old
who had suffered from epilepsy since childhood. Point C
in the temporal lobe was excited five times at intervals of
from five to 10 minutes, and after each stimulation the
patient's mood became friendlier; she smiled, questioned the
therapist directly about his nationality, background, and
friends, and declared that he "was nice," that his country
(Spain) "must be very beautiful," that "Spaniards are very
attractive," and she ended with the statement "I would like
to marry a Spaniard." This particular train of thought and
manner of speaking seemed completely spontaneous, but it
appeared only after stimulation of point C in the temporal
lobe, and no such shift to a flirtatious mood was noted
in her spontaneous conversations following stimulations of
other cerebral points.

The third case of evoked change in sexual ideology was

a young epileptic boy, A. F., who, following stimulation of
point LP 5-6 in the left temporal cortex, suddenly began
to discuss his desire to get married. After subsequent stimu-
lations of the point, he elaborated on this subject, revealed
doubts about his sexual identity, and voiced a thinly veiled
wish to marry the male interviewer.

Experiential Hallucinations

Hallucinations evoked by electrical stimulation of the
brain have been lucidly described by Roberts ( 1961 ), who
wrote: "It is as though a wire recorder, or a strip of cine-
matographic film with sound track, had been set in motion
within the brain. A previous experience — its sights and
sounds and the thoughts — seems to pass through the mind
of the patient on the operating table. ... At the same time
he is conscious of the present. . . . The recollection of the
experiential sequence stops suddenly when the electric cur-
rent ceases. But it can again be activated with reapplication
of the electric current." The hallucination may develop
during the stimulation, with a normal-like progression of
movements and sounds, which appear more real and vivid
than when the events actually happened. It is as if the pa-
tient had a double life, one in the past recalled by the elec-
trical stimulation, and another in the present, perceiving all
the sensory stimulation of the surroundings, but both with
a similar quality of reality, as if the person had a "double
consciousness" of subjective sensations. In some cases, com-
ponents of the hallucination are completely new and do
not belong to the subject's past experience, but usually, as
Penfield (1952, 1958, 1960) emphasized, the responses
are a detailed reenactment of previous experiences, an exact
"flash-back" activation of memories.

In one of our patients with intracerebral electrodes, de-
tailed study of the tape-recorded interviews demonstrated
that the perceptual content of some experiential responses

was related to the patient's thoughts at the moment of
stimulation. For example, when the patient was talking
about her daughter's desire for a baby sister, a stimulation
was applied to the temporal lobe and the patient heard a
female voice saying "I got a baby — sister." Baldwin ( 1960)
has reported a similar observation in which the content of
visual hallucinatory responses evoked in a 28-year old man
varied with the sex and identity of the observer seated be-
fore him in the operating room. In a previous article (Mahl
et at., 1964) we have suggested that "The patient's 'mental
content' at the time of stimulation is a determinant of the
content of the resulting hallucinatory experiences," and we
offered the so-called "altered-state hypothesis" in which the
essential effect of stimulation is to alter the state of con-
sciousness of the patient in such a way that primary proc-
ess thinking replaces secondary process thinking. (See
Freud, 1900.) According to this hypothesis, the electrical
stimulation of the temporal lobe would not activate memory
traces in the ganglionic record, as postulated by Penfield,
but would induce a state of consciousness which would in-
crease the functional probability of primary processes.

Pleasure

The possibility that "pleasure centers" might exist in the
brain was supported by the extensive work of Olds and his
collaborators (1954, 1956, 1961), who demonstrated that
rats prefer to stimulate some points of their brains by press-
ing a treadle, than to satisfy drives of hunger, thirst, and
sex. Positive behavioral qualities of cerebral stimulation
have been confirmed in other species including the cat (Sid-
man et al., 1955) and the monkey (Bursten and Delgado,
1958). However, "pleasure" has an experiential factor
which animals cannot report because they lack verbal com-
munication. Only studies in humans could reveal whether

electrical stimulation of the brain is able to induce pleasur-
able sensations. The study of patients with implanted elec-
trodes yielded affirmative evidence (Delgado, 1960; Sem-
Jacobsen and Torkildsen, 1960). In one of our cases,
stimulation of the temporal lobe evoked "pleasant tingling
sensations of the body" which were openly declared to be
very enjoyable. The patient's mood changed from its usual
peaceful state to one of giggling and laughing. She teased
the doctor and made fun of the experimental situation with
humorous comments.

In another patient, temporal-lobe stimulation evoked
"statements avowing his pleasure at being 'up here' and
'subject to us' which were classified as 'passive compliance' '
(Higgins et al., 1956). For example, when the patient
had been silent for five minutes, a point in the temporal
cortex was stimulated and he immediately exclaimed, "Hey!
You can keep me longer here when you give me these; I like
those." and he insisted that the "brain wave" testing made
him "feel O.K." Similar statements followed stimulation of
other temporal points, but were never expressed spontane-
ously in the absence of excitations. The statistical signifi-
cance of these results was P <0.001. as contrasted by X-
analysis.

During increased pleasure, the subjects were oriented
mainly toward themselves, and they often reported experi-
encing agreeable physical sensations, while during artifici-
ally increased speech and changes in sexual ideology they
expressed friendliness for the nearby people. In both cases,
there was a shift of emotional mood to a happy interpreta-
tion of reality, and this experience was interpreted by the
patient as spontaneous and valid, usually without being
directly related to the stimulation. A shift from pleasurable
thinking to friendliness and to sexual ideas has been ob-
served in some cases.

CONSFQUFNC ES OF BRAIN CONTROL

Probably the most significant conclusion derived from
electrical stimulation of the awake brain is that functions
traditionally related to the psyche such as friendliness,
pleasure, and verbal expression can be induced, modified,
and inhibited by direct stimulation of cerebral structures.
This discovery may be compared with the revolutionary
tinding almost two centuries ago that contraction of frog
muscle may be induced by electricity without need of the
soul's "animal spirits," because experimental analysis of
mental functions can now proceed without implicating
metaphysical entities. Research concerning the electrical
driving of emotions, anatomical correlates of memory, or
electrical signals related to learning does not interfere with
personal ideas about the natural or supernatural destiny of
man and does not involve theological questions, which
should be disassociated from neurophysiological inquiry. In
addition to electrical stimulation, there are now techniques
for exploration of brain function which include electrical
recording, chemical stimulation, intracerebral chemistry,
and electron microscopy. The task that we are facing is the
correlation of neuro-anatomy and physiology with mental
functions; the investigation of cerebral areas involved in
psychic manifestations; the analysis of their electrical and
chemical background; and the development of methods to
induce or inhibit specific activities of the mind.

Already we know that some structures, including the
hypothalamus, amygdala, central gray, and temporal lobe,
are involved in emotional phenomena, while other areas,
such as the parietal cortex, do not seem to participate in
psychic experience. Brain research has expanded rapidly
in recent years with the creation of institutes for multi-
disciplinary studies, but this field should attract even more
of our intellectual and economic resources. Human behav-

ior, happiness, good, and evil are, after all, products of
cerebral physiology. In my opinion, it is necessary to shift
the center of scientific research from the study and control
of natural elements to the analysis and patterning of mental
activities. There is a sense of urgency in this redirection
because the most important problem of our present age is
the reorganization of man's social relations. While the mind
of future generations will be formed by pedagogic, cultural,
political, and philosophical factors, it is also true that edu-
cation is based on the transmission of behavioral, emotional,
and intellectual patterns related to still unknown neuro-
physiological mechanisms. Investigators will not be able to
prevent the clash of conflicting desires or ideologies, but
they can discover the neuronal mechanisms of anger, hate,
aggressiveness, or territoriality, providing clues for the di-
rection of emotions and for the education of more sociable
and less cruel human beings. The precarious race between
intelligent brains and unchained atoms must be won if the
human race is going to survive, and learning the biological
mechanisms of social relations will favor the cerebral
victory.

Electrical and chemical analyses of mental functions
have introduced new facts into the much debated problem
of mind-brain relations. In the interpretation of data, we
should remember that spike potentials, neurohumors, and
synaptic transmitters may represent happiness and sorrow,
love and hate, war and peace, and in the near future we
can expect to find answers to classical questions concerning
psychological aspects of the physical brain. How can elec-
trical stimulation of the temporal lobe be felt as pleasure,
music, or fear? Why is a ferocious monkey tamed by apply-
ing a few volts of electricity to its caudate nucleus? As
discussed in a previous article (Delgado, 1964b), psycho-
physical correlations may be related to the two elements

which transmit information in the nervous system, namely,
the material carrier and the symbolic meaning. In the re-
ception of sensory inputs, there is an initial electrical cod'
ing which is the carrier necessary for neural circulation of
impulses. When a monkey, a savage, or a civilized man
looks at a pencil, the received visual stimulus is transformed
into electrical signals and transmitted through optic path-
ways to the brain. At the levels of retina and optic nerve, the
coding of the stimulus depends on the visual input, inde-
pendent of its possible meaning. Symbolism is created by
the association within the brain of two or more sensory
receptions or of present and past experiences, but it does
not depend on the material structure of the object or on
the pattern of its electrical coding. For a naive monkey or
for a savage, the pencil is a neutral object; for a writer, the
pencil is full of associations, uses, and meaning. Symbolism
is not intrinsic in the object, nor inborn in the brain: it
must be learned. The most important symbolic tool of the
mind, language, is not invented by each individual; it is a
cultural gift of the species. The symbolic meaning may be
considered an immaterial element of mental functions in
the sense that it is related to a spatio-temporal association
between two or more sensory receptions and not to the
material structure of the inputs. The elements for symbolic
recognition already exist in the electrical code of the trans-
mitted signals; however, they are not determined by the
pattern of the code but by spatio-temporal relations between
present and past codes which cannot be deciphered by any
instrument if the reference point of the past is not known.
These temporal and spatial relations may be considered
as material or immaterial, depending on the investigator's
point of view. Obviously, the relations depend on the mate-
rial existence of some events, but, at the same time, the
relations are independent of the material organization of

each event. It is a question of definition, and, if we explain
the meaning of our terms, there is no conflict. I think, how-
ever, that it is more practical to consider symbolism as
non material in order to emphasize the relativity of its
existence and the fact that it does not depend on the intrin-
sic qualities of matter but on the previous history of the
object and of the observer. In the last analysis, behavior
could be reduced to movement of atoms, but if we are dis-
cussing the emotional behavior of the monkey, it would be
difficult to explain it in terms of orbiting particles, and it is
far more useful to employ psychological concepts. It should
be clarified that, in the observer, conscious understanding
of meaning is probably dependent upon progressive steps
of electrical subcoding of sensory inputs with the creation
of new material and symbolic elements related to the activa-
tion of a new series of chemical and electrical phenomena
affecting specialized neurons. However, the distinction be-
tween material carrier and symbolic meaning simplifies the
interpretation of neurophysiological data, because analysis
of events in receptors and in transmitting pathways will
provide information about the carrier but not about sym-
bols. At the same time, it should be expected that electrical
stimulation of neuronal groups may activate processes re-
lated to both material carriers and symbolic meaning. This
working hypothesis may help in the differentiation between
cerebral mechanisms responsible for transmitting inputs
and for cognitive processes of received signals.

From its beginning, wiring of the human brain aroused
emotional opposition even among scientists, while similar
wiring of the heart or of the bladder has been received
enthusiastically. The difference in attitude was no doubt
related to a more or less conscious personal fear that our
identity could be attacked and that our mind could be
controlled. Personal traits such as friendliness, sexual in-

clination, or hostility have already been modified during
cerebral stimulation, and we can foresee other influences
on emotional tone and behavioral reactions. Eleetrieity is
only a trigger of pre-existing mechanisms which could not,
for example, teach a person to speak Spanish, although it
could arouse memories expressed in Spanish if they were
already stored in the brain.

Entering into the field of speculation, I would like to
comment on one question which has already caused wide-
spread concern. Would it be feasible to control the behavior
of a population by electrical stimulation of the brain? From
the times of slavery and galleys up to the present forced-
labor camps, man has certainly tried to control the behavior
of other human beings. In civilized life, the intervention of
governments in our private biology has become so deeply
rooted that in general we are not aware of it. Many coun-
tries, including the United States, do not allow a bride and
groom to marry until blood has been drawn from their veins
to prove the absence of syphilis. To cross international
borders, it is necessary to certify that a scarification has
been made on the skin and inoculated with smallpox. In
many cities, the drinking water contains fluoride to
strengthen our teeth, and table salt is fortified with iodine
to prevent thyroid misfunction. These intrusions into our
private blood, teeth, and glands are accepted, practised,
and enforced. Naturally, they have been legally introduced,
are useful for the prevention of illness, and do generally
benefit society and individuals, but they have established a
precedent of official manipulation of our personal biology,
introducing the possibility that governments could try to
control general behavior or to increase the happiness of
citizens by electrically influencing their brains. Fortunately,
this prospect is remote, if not impossible, not only for obvi-
ous ethical reasons, but also because of its impracticability.

Theoretically it would be possible to regulate aggressive-
ness, productivity, or sleep by means of electrodes im-
planted in the brain, but this technique requires specialized
knowledge, refined skills, and a detailed and complex ex-
ploration in each individual, because of the existence of
anatomical and physiological variability. The feasibility of
mass control of behavior by brain stimulation is very un-
likely, and the application of intracerebral electrodes in
man will probably remain highly individualized and re-
stricted to medical practice. Clinical usefulness of electrode
implantation in epilepsy and involuntary movements has
already been proved, and its therapeutical extension to
behavioral disorders, anxiety, depression, and other illness
is at present being explored. The increasing capacity to
understand and manipulate mental functions of patients
will certainly increase man's ability to influence the be-
havior of man.

If we discover the cerebral basis of anxiety, pleasure,
aggression, and other mental functions, we shall be in a
much better position to influence their development and
manifestations through electrical stimulation, drugs, sur-
gery, and especially by means of more scientifically pro-
grammed education.

These possibilities pose tremendous problems. As Skin-
ner asked recently (1961), "Is the deliberate manipulation
of a culture a threat to the very essence of man or, at the
other extreme, an unfathomed source of strength for the
culture which encourages it?" Scientific discoveries and
technology cannot be shelved because of real or imaginary
dangers, and it may certainly be predicted that the evolu-
tion of physical control of the brain and the acquisition of
knowledge derived from it will continue at an accelerated
pace, pointing hopefully toward the development of a more
intelligent and peaceful mind of the species without loss of

individual identity, and toward the exploitation of the mosl
suitable kind oi feedback mechanism: the human brain

studying the human brain.

ACKNOWLEDGMENTS

Part of the researeh mentioned in this paper was sup-
ported by grants from the United States Public Health
Service and the Office of Naval Research. Some of the
studies were conducted during a John Simon Guggenheim
fellowship. The experimental and editorial collaboration of
Caroline Delgado is warmly acknowledged.

REFERENCES CITED

Baldwin, M.

1960. Electrical stimulation of the mesial temporal region. //; Ramey, E. R.,
and D. S. O'Doherty (eds.). Electrical studies on the unanesthetized
brain. New York. Paul B. Hoeber. pp. 159-176.

BlJRSTEN, B., AND J. M. R. DELGADO

1958. Positive reinforcement induced by intracerebral stimulation in the
monkey. Jour. Comp. Physiol. Psychol., vol. 51, pp. 6-10.

Delgado, J. M. R.

1955a. Cerebral structures involved in transmission and elaboration of noxious

stimulation. Jour. Neurophysiol., vol. 18, pp. 261-275.
1955b. Evaluation of permanent implantation of electrodes within the brain.

EEG Clin. Neurophysiol., vol. 7, pp. 637-644.

1959. Prolonged stimulation of brain in awake monkeys. Jour. Neurophysiol..
vol. 22, pp. 458-475.

1960. Emotional behavior in animals and humans. Psychiat. Res. Rept., vol.
12, pp. 259-271.

1962. Pharmacological modifications of social behavior. //; Paton, W. D. M.
(ed.). Pharmacological analysis of central nervous action. Oxford, Per-
gamon Press, pp. 265-292.

1963a. Effect of brain stimulation on task-free situations. EEG Clin. Neuro-
physiol., Suppl. 24, pp. 260-280.

1963b. Telemetry and telestimulation of the brain. In Slater, L., (ed.), Bio-
telemetry. New York, Pergamon Press, pp. 231-249.

1963c. Cerebral heterostimulation in monkey colony. Science, vol. 141, pp.
161-163.

1964a. Free behavior and brain stimulation. //; Pfeiffer, C. C, and J. R.
Smythies (eds.), International review of neurobiology. New York,
Academic Press, vol. 6, pp. 349-449.

1964b. Factores extracerebrales de la mente. Rev. Occidente, no. 14, pp.
131-144.
Delgado. J. M. R., F. J. Castejon y F. Santisteban

1964. Radioestimulation cerebral en toros de lidia. VIII Reun. Nac. Soc.
Ciencias Fisiologicas, Madrid, Febrero.
Ewald, J. R.

1898. Ueber kiinstlich erzeugte Epilepsie. Berliner Klin. Wochenschr., vol.
35, p. 689.
Feindel, W.

1961. Response patterns elicited from the amygdala and deep temporo-insular
cortex. In Sheer, D. E. (ed.), Electrical stimulation of the brain. Austin,
Texas, University of Texas Press, pp. 519-532.

FONBERG, E., AND J. M. R. DELGADO

1961. Avoidance and alimentary reactions during amygdala stimulation. Jour.
Neurophysiol., vol. 24, pp. 651-664.

1 KM I). S.

1900. The interpretation o( dreams, Standard edition ol complete psycho-
logical works of Sigmund Freud. London, Hogarth Press, 1953. vols.
4. 5.
t k 1 1 s( ii. Ci.. and K. Hitzig

1870. Uc her die eleklrische I i rcgh;u kcit des ( irosshirns. Arch. \nat. Physiol.,
1 eipzig, vol. 37. pp. 300-332.
(ii i I HORN, E.

1957. Autonomic imbalance and the hypothalamus. Implications for physi-
ology, medicine, psychology and neuropsychiatry. Minneapolis, Uni-
versity of Minnesota Press. 300 pp.
Hi mh. R. G.

1954. Studies in schizophrenia. A multidisciplinary approach to mind-brain
relationships. Cambridge. Harvard University Press, 619 pp.
Hiss. W. R.

1932. Beitrage zur Physiologie d. Hirnstammes I. Die Methodik der lokali-
sierten Reizung und Ausschaltung subkortikaler Hirnabschnitte. Leip-
zig, Georg Thieme, 122 pp.
1954. Diencephalon. Autonomic and extrapyramidal functions. New York,
Grune and Stratton, 79 pp.
Higgins. J. W., G. F. Mahl, J. M. R. Delgado, and H. Hamlin

1956. Behavioral changes during intracerebral electrical stimulation. Arch.
Neurol. Psychiat., Chicago, vol. 76, pp. 399-419.

H()DLS, R., AND H. W. MAGOUN

1942. Autonomic responses to electrical stimulation of the forebrain and
midbrain with special reference to the pupil. Jour. Comp. Neurol.,
vol. 76, pp. 169-190.
Kwda.B. R.

1951. Somato-motor, autonomic and electrocorticographic responses to elec-
trical stimulation of "rhinencephalic" and other structures in primates,
cat and dog. Acta Physiol. Scandinavica, vol. 24, Suppl. 83, 285 pp.
King. H. E.

1961. Psychological effects of excitation in the limbic system. In Sheer, D. E.
(ed.), Electrical stimulation of the brain. Austin. Texas, University of
Texas Press, pp. 477-486.
Mahl. G. F., A. Rothenberg, J. M. R. Delgado, and H. Hamlin

1964. Psychological responses in the human to intracerebral electric stimula-
tion. Psychosom. Med., vol. 26. pp. 337-368.
Olds, J.

1956. Pleasure centers in the brain. Sci. Amer.. vol. 195, pp. 105-1 16.
1961. Differential effects of drives and drugs on self-stimulation at different
brain sites. //; Sheer, D. E. (ed.). Electrical stimulation of the brain,
Austin. Texas, University of Texas Press, pp. 350-366.
Olds, J., and P. Milner

1954. Positive reinforcement produced by electrical stimulation of the septal
area and other regions of the rat brain. Jour. Comp. Physiol. Psychol.,
vol. 47. pp. 417-428.

Penfield, W.

1952. Memory mechanisms. Arch. Neurol. Psychlat., Chicago, vol. 67, pp.

178-198.
1958. The excitable cortex in conscious man. The Sherrington Lectures V.

Springfield. Illinois, C. C. Thomas.

1960. A surgeon's chance encounter with mechanisms related to conscious-
ness. Jour. Roy. College Surgeons Edbinburgh, vol. 5, p. 173.

Penfield, W., and H. Jasper

1954. Epilepsy and the functional anatomy of the human brain. Boston,
Little Brown, 896 pp.

Roberts, L.

1961. Activation and interference of cortical functions. //; Sheer, D. E. (ed.),
Electrical stimulation of the brain. Austin, Texas, University of Texas
Press, pp. 533-553.

Sem-Jacobsen, C. W., M. C. Petersen, H. W. Dodge, Jr., J. A. Lazarte, and

C. B. HOLMAN

1956. Electroencephalograph^ rhythms from the depths of the parietal,
occipital and temporal lobes in man. EEG Clin. Neurophysiol., vol. 8,
pp. 263-278.
Sem-Jacobsen, C. W., and A. Torkildsen

1960. Depth recording and electrical stimulation in the human brain. //;
Ramey, E. R., and D. S. O'Doherty (eds.), Electrical studies on the
unanesthetized brain. New York, Paul. B. Hoeber, pp. 275-290.

Showers, M. J. C, and E. C. Crosby

1958. Somatic and visceral responses from the cingulate gyrus. Neurology,
vol. 8. pp. 561-565.
Sidman, M., J. V. Brady, J. J. Boren, D. G. Conrad, and A. Schulman

1955. Reward schedules and behavior maintained by intracranial self-stimula-
tion. Science, vol. 122, pp. 830-831.

Skinner, B. F.

1961. The design of cultures. Daedalus, pp. 534-546.
Walker, A. E., and C. Marshall

1961. Stimulation and depth recording in man. In Sheer, D. E. (ed.), Elec-
trical stimulation of the brain. Austin, Texas, University of Texas
Press, pp. 498-518.

JAMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

1968

THREE VIEWS OF THE
NERVOUS SYSTEM

KENNETH D. ROEDER

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1968

JAMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

JAMES ARTHUR LECTURE ON

THE EVOLUTION OF THE HUMAN BK\I\

1 9ft S

THREE VIEWS OF THE
NERVOUS SYSTEM

KENNETH D. ROEDER

Professor of Physiology
Department of Biology

Tufts University
Medford, Massachusetts

THE AMERICAN MUSEUM OF NATURAL HISTORY-
NEW YORK : 1968

JAMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

l rederick Tilney, The Brain in Relation to Behavior; March 15, 1932

C. Judson Herrick, Brains as Instruments of Biological Values; April ft. 1933

D. M. S. Watson. The Story of Fossil Brains from Fish to Man; April 24, 1934

C. U. Aricns Rappers. Structural Principles in the Nervous System; The Develop-
ment of the Forebrain in Animals and Prehistoric Human Races; April 25.
1935

Samuel T. Orton. The Language Area of the Human Brain and Some of its Dis-
orders; May 15. 1936

R. W. Gerard, Dynamic Neural Patterns; April 15, 1937

Franz Weidenreich, The Phylogenetic Development of the Hominid Brain and its
Connection with the Transformation of the Skull; May 5, 1938

G. Kingsley Noble, The Neural Basis of Social Behavior of Vertebrates; May 1 1,
1939

John F. Fulton, A Functional Approach to the Evolution of the Primate Brain;
May 2. 1940

Frank A. Beach, Central Nervous Mechanisms Involved in the Reproductive Be-
havior of Vertebrates; May 8, 1941

George Pinkley, A History of the Human Brain; May 14, 1942

James W. Papez. Ancient Landmarks of the Human Brain and Their Origin;
May 27, 1943

James Howard McGregor, The Brain of Primates; May 1 1, 1944

K. S. Lashley, Neural Correlates of Intellect; April 30, 1945

Warren S. McCuIloch, Finality and Form in Nervous Activity; May 2, 1946

S. R. Detwiler, Structure-Function Correlations in the Developing Nervous Sys-
tem as Studied by Experimental Methods; May 8, 1947

Tilly Edinger. The Evolution of the Brain; May 20, 1948

Donald O. Hebb, Evolution of Thought and Emotion; April 20, 1949

Ward Campbell Halstead, Brain and Intelligence; April 26, 1950

Harry F. Harlow, The Brain and Learned Behavior; May 10, 1 95 1
Clinton N. Woolsey, Sensory and Motor Systems of the Cerebral Cortex; May 7,
1952

Alfred S. Romer, Brain Evolution in the Light of Vertebrate History; May 21,
1953

Horace W. Magoun, Regulatory Functions of the Brain Stem; May 5, 1954

Fred A. Mettler, Culture and the Structural Evolution of the Neural System;
April 21, 1955

Pinckney J. Harman, Paleoneurologic, Neoneurologic, and Ontogenetic Aspects
of Brain Phylogeny; April 26, 1956

Davenport Hooker, Evidence of Prenatal Function of the Central Nervous System
in Man; April 25, 1957

David P. C. Lloyd, The Discrete and the Diffuse in Nervous Action; May 8, 1958

Charles R. Noback, The Heritage of the Human Brain; May 6, 1959

Ernst Scharrer, Brain Function and the Evolution of Cerebral Vascularization;
May 26, 1960

Paul I. Yakovlev, Brain, Body and Behavior. Stereodynamic Organization of the
Brain and of the Motility -Experience in Man Envisaged as a Biological
Action System; May 16, 1961

H. K. Hartline, Principles of Neural Interaction in the Retina; May 29, 1962

Harry Grundfest, Specialization and Evolution of Bioelectric Activity; May 28,
1963

Roger W. Sperry, Problems Outstanding in the Evolution of Brain Function;
June 3, 1964

Jose M. R. Delgado, Evolution of Physical Control of the Brain; May 6, 1965

Seymour S. Kety, Adaptive Functions and the Biochemistry of the Brain; May 19,
1966

Dominick P. Purpura, Ontogenesis of Neuronal Organizations in the Mammalian
Brain; May 25. 1967

Kenneth D. Roeder, Three Views of the Nervous System; April 2, 1968

THREE VIEWS OF THE
NERVOUS SYSTEM

INTRODUCTION

The theme of the James Arthur Lecture series is the
evolution of the human brain. Taken in its broadest func-
tional sense, this topic is the most baffling that faces
biology today, for man is trying to understand the instru-
ment of his own intelligence. Part of the problem is that
there is at present no hint of a "break-through" — nothing
equivalent to the elucidation of the DNA structure that led
to such an upsurge of work on the mechanisms of inheri-
tance. The only recourse is attacks on the problem from
many directions, some seemingly oblique and indirect. One
of these directions seeks to understand how simpler nervous
systems determine adaptive behavior.

This may be taken as a formal justification for my
presence here, although the truth is that I work with insect
nervous systems because I enjoy it. But one cannot remain
absorbed in any specialty for thirty years without won-
dering about its wider implications. Therefore, I welcome
this chance to discuss the working of the nervous systems
of insects in relation to the way insects behave, and to
search for a view of my interest set in a wider framework.

INSECTS AND VERTEBRATES

Insects and vertebrates compete ecologically to a degree
found in no other two classes of land animals. Man, as the
dominant vertebrate, bears the brunt of this competition.
There are mutterings in some quarters about the advantage
to man in the extermination of this or that insect species

or even of whole insect groups. To my mind it makes better
biological sense to compare the workings of our competitors
with our own with the object of outmaneuvering them
rather than exterminating them.

Insects and vertebrates represent widely divergent
branches of the phylogenetic tree. Consequently, they show
striking contrasts as well as similarities. Because these con-
trasts and similarities are important to my general theme,
I shall begin by commenting briefly on examples of each.

Some of the contrasts are self-evident. Approximately
one million insect species have been described, and it is
estimated that millions more await description. Approxi-
mately thirty thousand vertebrate species have been cata-
logued. Individuals of the great majority of insect species
weigh less than one-tenth of an ounce; some vertebrates
weigh many tons. This is not the place to discuss the archi-
tectural plan of the insect skeleton and how it has imposed
a mechanical upper limit on its body size. An important
corollary of this size limitation, however, is that insect
nervous systems are correspondingly small, even though
some of their neurons are as large as or larger than our
own. It follows that insect nervous systems must contain
fewer neurons, and that there must be parsimony in the
way neurons are involved in the multifarious patterns of
insect behavior. I shall try to illustrate this at a later point.

Insect and vertebrate similarities are, at first glance, less
apparent. It is generally true, however, that if one dissects
different animals and inspects their body mechanisms, the
similarities become more apparent as the grain of the in-
spection becomes finer. For instance, at the molecular level
nearly all living things find a common ground. At a coarser
level, say, that of the light microscope, it is still much
easier to determine by inspection what the tissues are for,
that is, contraction, conduction, or secretion, than it is to

say whether they belong to an insect or to a vertebrate.
This is also true when such tissues are functionally exam-
ined. For instance, insect neurons and vertebrate neurons
seem to operate on the same general principles.

THREE VIEWPOINTS

Comparing the workings of insect and human brains is
like trying to understand a strange and primitive culture
from the viewpoint of our own civilization. The outward
cultural expressions — mores, economics, religion, and
"foreign policy" — seem to us quite difficult to understand,
and we can make only blanket generalizations from an
external study. On learning more about individual mem-
bers of that culture, we find that they are very like our-
selves and that they have the same joys, anxieties, and
motivations. The last and most difficult stage of under-
standing is to learn how individual members of the citizenry
relate to their fellows to form the cultural mesh that deter-
mines the image of the strange land.

I shall try to present what I know about insect brains
and behavior from three similar viewpoints. First I shall
discuss in a general fashion the functions of the insect
brain in relation to certain behavioral patterns. Next, I
shall summarize the main attributes of that common
denominator of all higher nervous systems, the neuron.
Finally, I shall attempt the most difficult task of all — to
examine how neurons transpose signals from the outer
world and interact with other neurons forming the neural
mesh to generate an adaptive behavioral pattern.

THE INSECT BRAIN

For an overview of any nervous system, it is best to
begin by glancing at its origins. Insect ancestors were
probably wormlike forms having a series of similar body

segments (fig. 1A). The activities of each body segment
were largely autonomous and were controlled by a ganglion
or, rather, a bilateral ganglion pair. The ganglia were
serially connected by a pair of longitudinal bundles of
nerve fibers. These connectives played little part in deter-
mining the local affairs of the individual segments and

Fig. 1. A. Nervous system of hypothetical ancestor of segmented worms.
B. Later stage in the evolution of the arthropod brain. The three anterior ganglia
have moved to a dorsal position and have become practically fused. C. Nervous
system of the praying mantis. A relatively unspecialized insect nervous system
with most of the ventral ganglia distinctly separated. The front pair of legs are
specialized for grasping prey.

served mainly to coordinate rapid movements such as
those needed in evading a predator. The system of the worm
can be likened to a group of self-sufficient rural communi-
ties that resort to cooperation only when faced by a general
threat.

At least three, and possibly more, of these ganglia lay
in front of the ventrally placed mouth of the worm. The
remainder were arrayed behind it and along the ventral
surface of the segment chain. The mouth moved to the front
end of this primitive creature (the logical spot for gathering
food), and the anterior ganglia came to assume a dorsal
position while fusing to form a brain (fig. IB). As the worm
became more mobile its "distance"' receptors, vision and
chemo-reception, clustered at its front end and became
more complex and discriminating. Neurons subserving
them multiplied, forming the bulk of the adjacent brain.

Broadly speaking, the nervous systems of insects still
follow this plan (fig. 1C). The organs of vision and olfac-
tion have increased enormously in complexity and diversity,
and corresponding regions of the brain have enlarged ac-
cordingly. Similarly, many of the body segments and their
appendages have diversified for walking, grasping, hopping,
swimming, flying, egg laying, and copulating. Others have
atrophied or become fused with their neighbors. The seg-
mental ganglia, however, still retain much of their primitive
autonomy in coordinating and regulating the local muscle
sequences needed for these special action patterns. The
brain plays no part in determining which, or in what order,
the muscles of a given segment will contract in performing
a given action. This is determined by the relevant seg-
mental ganglion or by the ganglia of a few adjacent seg-
ments acting in concert, as in the coordination of the three
pairs of legs during walking.

CONTROL BY INHIBITION

At first glance, this arrangement seems to leave the
brain with no higher function beyond that needed to process
the information coming from the eyes and antennae. There
is much evidence, however, that the brain exerts what might

be called an over-all direction, or command function, in
determining the particular action pattern or behavior mode
shown by the whole insect under a particular set of condi-
tions. This control seems to be exerted primarily through
selective suppression of certain of the locally organized
action patterns, the behavioral mode shown at any given
moment being released from this suppression. This conclu-
sion is based on experiments such as the following.

The praying mantis waits in ambush for its food, and
thus remains motionless most of the time; after removal of
its brain a mantis walks continuously (Roeder, 1937).
Most insects exhibit sexual behavior only in the presence
of appropriate releasing stimuli provided by the opposite
sex; decapitated male mantids make continuous copulatory
movements irrespective of the presence of a female
(Roeder, 1935). Ovipository behavior seems to be similarly
controlled. The motor patterns responsible for song pro-
duction in crickets are coordinated by the thoracic ganglia,
yet song patterns specifically connected with different court-
ship phases can be released in inappropriate circumstances
by electrical stimulation of certain regions of the brain
(Huber, 1960, 1967). Flapping of the wings in flight
normally ceases as soon as the feet of an insect touch the
ground. If the insect is, however, decapitated while in flight
this natural "either or" method of replacing the flight mode
by the walking mode is often ineffective, the insect continu-
ing in its attempt to fly even after tarsal contact has been
made.

These examples suggest that a considerable proportion
of the direction from the head ganglia is accomplished by
proscription, that is, by selective suppression of specific
activities generated and organized in the ganglia of the
several body segments. There is further evidence that inhi-
bition may occur at several levels within the brain. Centers

in the right and left halves having inhibitory control over
activities organized at a lower level may also inhibit one
another ( Roeder, 1937). Although the brain seems to have
this "either or" control over what the whole insect docs.
the same principle extends to the local segmental activities
presided over by the segmental ganglia. This is evident in
the control of alternate stepping movements of the right
and left legs of a segment and in the control of grooming
behavior in locusts (Rowell, 1965).

THE "ONENESS" OF BEHAVIOR

One of the most commonplace, but to me most remark-
able, aspects of the behavior of animals is the "oneness"
or singularity of their acts. An animal seems able to select
just one mode of behavior even under such circumstances
as being exposed to stimuli capable at other times of re-
leasing a wide variety of behavior patterns. It is easy to
justify the adaptive value of this unity of response, but,
regarded mechanistically, it seems surprising that a system
with so many input channels should so rarely compromise
between conflicting signals. In essence, this problem is one
of "attention." which is no less marked in insects than in
higher animals. It is also present at lower levels of the nerv-
ous system, for the reflex contraction of one muscle group
automatically inhibits the contraction of its antagonist
muscles.

Do the command functions of the insect brain play a
part in this "oneness" of behavior? In releasing one be-
havioral pattern does the brain increase the suppression
of others? If such is the case one would expect to observe
conflicting behavior patterns in a brainless insect.

There is some evidence for this. A praying mantis nor-
mally remains motionless for hours at a time, waiting in
ambush at the top of a vertical surface. From this vantage

point it strikes at passing insects which are grasped in its
specially modified forelegs. If placed on the ground a mantis
will usually walk until it encounters a vertical object, such
as a plant stem. It then climbs to the top of this object and
remains motionless in the in-ambush posture. After the
removal of its brain, however, a mantis walks continually,
persisting in its attempts to travel forward even after reach-
ing the top of a vertical object. If, during these travels, a
twig or other small object happens to touch the inner,
spined surface of its foreleg, the object is grasped firmly
and persistently. The insect appears to be unable to release
its grip even though this action may impede further for-
ward progress. The action of grasping does not, however,
suppress continuous attempts to walk forward, with the
result that the insect frequently becomes hopelessly en-
tangled in twigs and grass stems (Roeder, 1937). It might
be thought that this abnormal behavior is due to sensory
deprivation, but it is not produced by removal of the eyes,
optic ganglia, or antennae. In the intact insect the two
action patterns (grasping and walking) rarely, if ever,
occur simultaneously, and their simultaneous appearance
in the brainless mantis places the insect in a behavioral
cul-de-sac. This suggests that, when the brain is present,
either one, or neither, but never both, of these behavioral
modes is released.

ENDOGENOUS ACTIVITY

There is little detailed physiological information as to
how these segmentally determined activities are organized.
Nor do we understand the nature of the inhibition that
patterns them locally and controls them selectively from
the brain. In some cases inhibition appears to operate by
raising the threshold of a locally organized reflex response,
that is, by rendering it less likely to occur. This is seen in

the grasping reflex of the mantis described above and in
the grooming reflexes of locusts (Rowell, 1965). In other
cases the segmental neural systems seem to be intrinsically
unstable, that is, capable of endogenous generation of be-
havior patterns. Organized sequences of nerve impulses
are transmitted to appropriate muscles even after the
ganglion has been deprived of all sensory input. This has
been shown to be the case with copulatory movements
generated by the last abdominal ganglion of the male pray-
ing mantis (Roeder, Tozian, and Weiant, 1960) and with
wing flapping in locusts (Wilson, 1961, 1967).

It has long been known (Adrian, 1931) that insect
ganglia discharge patterns of impulses for considerable
periods after they have been isolated from all sensory input.
Some of this endogenous activity may be abnormal (Rowell,
1965), that is, caused by the surgical insult and unrelated
to normal behavior. In the two cases cited above, however,
endogenous neural activity seems to be the basis for move-
ments that have significance in the lives of the animals
concerned. There is, indeed, no satisfactory way to dis-
tinguish between a reflex response, the threshold of which
has been reduced to extremely low levels, and a system
that is endogenous or self -excitatory (Roeder, 1955).

THE BEHAVIOR OF NEURONS

So far, I have considered only the external or behavioral
signs of nervous system function. I have glanced, as it were,
at the "foreign policy" of the cell community that makes
up an animal. In the preceding paragraph it was necessary
to mention neurons and nerve impulses. Neurons are the
unit components of the nervous system or, if you prefer
the sociological analogy, members of the community that
formulate the foreign policy. Somehow, the details of the
mass transactions between brain and ventral ganglia must

originate in transactions between neurons. Such transac-
tions are accomplished mainly through nerve impulses.

It is perhaps as misleading to generalize about a "neuron"
as it is to generalize about a "person." The transactions
of neurons in the central nervous system have been most
closely scrutinized in studies of vertebrates, particularly
through the monumental work of Eccles (1953, 1964)
on the spinal cord of the cat. There is no evidence that
insect neurons operate on basically different principles,
so I shall draw largely on this work in making my brief
generalizations.

The central nervous system can be regarded as an or-
ganized mesh of nerve fibers. Extending into this mesh
are fibers from a multitude of sensory neurons (sense cells)
that are acted upon by the outer world. Out of the mesh
extend fibers belonging to motor neurons. These connect
with effectors — the muscle fibers and gland cells that act
upon the outer world. The patterning of muscle contrac-
tions that manifests itself as behavior is determined in part
by the organization and functional state of neurons forming
the central mesh and in part by the pattern of input signals
reaching the central mesh from sensory neurons.

Those neurons lying entirely within the central mesh
are called interneurons. They are of many sizes and config-
urations and have many ways of interacting. I must neglect
entirely the interactions based on neurosecretion and hor-
mones, and will limit this discussion to rapid, short-term,
neuron transactions carried out by means of nerve im-
pulses.

A generalized diagram of an insect interneuron is
shown in figure 2. It receives excitation from impulses
arriving at close contacts (synapses) after traveling in
nerve fibers (axons) belonging to other neurons. Nerve
impulses can be detected as small, transient, electrical

"spikes" propagating along a nerve fiber. Information is
contained in the frequency, timing, and pattern with which

nerve impulses reeur.

Fig. 2. A generalized diagram of an insect interneuron. At left, four pre-
synaptic fibers make two inhibitory (open circles) and two excitatory (solid
circles) contacts with its dendrites. The axon of the interneuron forms excitatory
synapses with other interneurons (at left). An electrode (arrow) leading to an
amplifier (A) registers the pattern of spikes discharged by the interneuron. When
presynaptic impulses arrive at two excitatory and one inhibitory synapse (upper
trace), the integrated result is an increase above the free-running spike frequency.
Activity of two inhibitory and one excitatory synapse (lower trace) causes a
decrease in spike frequency.

Synaptic contacts are of several kinds and are often
highly complex, but in the present context their most im-
portant property is that most of them represent a hiatus
or hindrance to the process of impulse propagation through
the mesh. This means that the arrival of an impulse at an

excitatory synapse does not generate in one-to-one fashion
another impulse in the downstream neuron. It merely in-
creases for a few milliseconds the tendency of the recipient
neuron to fire off an impulse of its own. The excitatory
state wanes exponentially. This means that impulses arriv-
ing roughly coincidentally at neighboring synapses formed
on the same interneuron will summate in promoting the
firing tendency of the neuron, which may cause it either
to discharge impulses or, if it is already active, to increase
its firing rate (fig. 2). In the same way, impulses arriving
with greater frequency at a given synapse summate in their
effects on the recipient neuron to a greater degree than if
they impinge on it at more extended intervals.

A proportion of the synapses formed on many inter-
neurons are inhibitory. The arrival of an impulse at an
inhibitory synapse decreases for a few milliseconds the
tendency of the recipient neuron to fire off an impulse of
its own. The collective effects of impulses arriving at several
inhibitory synapses summate in time and space as do those
arriving at excitatory synapses, and the effects of both
types are continuously integrated by the recipient inter-
neuron. Thus, one must picture an interneuron as being
exposed to a running barrage of excitatory and inhibitory
effects, each with a "half-life" of a few milliseconds. Its
own discharge pattern reflect the running integration of
this barrage. Elsewhere (Roeder, 1967b) I have compared
the activity of an interneuron to the actions taken by an
administrator. He bases his actions on decisions reached by
integrating the positive and negative opinions of others,
the most recent opinions being the most influential. Some
interneurons, like lower-level administrators, merely relay
forward the impulse pattern reaching their synapses. But
in the central nervous system these are probably in the
minority, and in any case their behavior is relatively unin-

teresting in our efforts to understand the transactions of
the brain.

References must be made to other sources (Eccles, I 953,
1964) for the details of synaptic action, chemical effects,
types of synapse, and the complex feedback arrangements
found in neuron populations. The point is that synaptic
interaction of neurons is the only known way in which
fast-acting integrations and transformations of the central
nervous system are carried out. Admittedly, it is hard to
believe that higher nervous functions, such as learning,
memory, and abstract thought, are based only on such a
system. New modes of neuron interaction and special
properties that emerge from the mesh may be discovered,
but it must be realized that it would be hard to predict the
properties of a computer if one were given only the prop-
erties of a single transistor.

Next, I shall attempt to describe some of the neuron
signals and transactions concerned in a relatively simple
piece of insect behavior.

MOTHS AND BATS

It is observed that a certain pattern of stimuli impinging
on an animal bears a causal relation to action having
adaptive value. The problem facing the neurophysiologist
is to untangle the mechanisms transforming stimulus into
action. Commonly, the problem is formidable at the outset;
the stimulus pattern may be complex and hence difficult
to define in physical or chemical terms, and it usually im-
pinges on the animal via thousands of receptor neurons,
each having a separate fiber leading to the central nervous
system. Therefore, many pathways must be monitored
simultaneously in order to assess fully the incoming sensory
information. The initial difficulty is often insurmountable,
but it must be overcome before one can know, in terms

Fig. 3. A, B. Photographic tracks registered by moths flying free in the field
at night. The loudspeaker on the mast began emitting a series of ultrasonic pulses
in simulation of a bat at the instant indicated by the bright spot in A and by the
arrow in B. The tracks have breaks every 0.25 second. The oscillations on the
tracks are due to the flapping wings of the moths. A. Diving in response to a
loud sound. B. Turning-away in response to a faint sound (Roeder, 1962).
B, C. Electronic registration of the attempts of a captive moth to turn away
from a loudspeaker emitting a train of faint ultrasonic pulses. Upward deflection
(top trace) indicates an attempt to make a right turn; downward deflection, an
attempt to turn left. Middle trace shows wing movemens of moth. Lower trace
indicates onset of pulse train (10 per second). Vertical grid marks 100-milli-
second intervals. C. Loudspeaker was in horizontal plane and at 90 degrees to
body axis of moth on left side. D. Same, loudspeaker on right side. Attempts to
turn away began about 50 milliseconds after first sound pulse. The moth was a
female of Leucania commoides (Roeder, 1967 ).

of nerve impulse patterns, how the outer world is being
reported to the central nervous system under the given

circumstances. The example I wish to present overcomes
this initial difficulty.

Several species of insectivorous bats of North America
h\ and teed in darkness. They use a kind of sonar to avoid
obstacles in their path and to find, track, and capture flying
insects. The operation of this sonar has been clarified by
the elegant work of Griffin and his students (1958). A
cruising bat emits a series of ultrasonic cries and appears
to be able to estimate the distance and direction of objects
in its flight path from changes in the echoes returning to

R 4 3

^ r ■ — u Y i

Sl ^2JR

wwvwvwwwwwvaai

Fig. 4. A. Ultrasonic cries recorded from a cruising bat thing in the field.
Time. 100 cycles per second. B. A Single cry on expanded time base. Time.
1000 cycles per second. C. Artificial ultrasonic pulse similar to those used in the
experiments described in the text. Vertical grid. 2-milliseconds per division.
D. Diagram showing the individually variable parameters of the stimulus: 1,
frequency: 2. amplitude: 3. pulse duration: 4. pulse repetition rate: 5. pulse
train length.

its ears. The range of this sonar system for an object the
size of a flying moth appears to be less than 10 feet.

Moths of several families, notably the Noctuidae. have
auditory organs maximally sensitive to the pitch of bat
cries. They serve the moth as counter-sonar detectors, and
they are able to register bat cries at distances of up to 130

feet (Roeder, 1966a). Moths show two types of reaction
when they are exposed to real or simulated bat cries (fig.
3 A, B). If the sounds reach the moths at high intensity,
as from a nearby bat, the insects show various kinds of
unpredictable behavior, such as twisting, turning, and
diving toward the ground. If the sounds are received at
low intensity, for example, by a moth flying 50 to 100 feet
distant from a hunting bat, the moth turns and steers a
course directly away from the source of the ultrasonic
pulses (Roeder, 1962).

The survival value of turning-away behavior is fairly
clear. It carries the slower-flying moth out of the feeding
area of the bat before its presence has been detected by the
sonar of the predator. Turning-away behavior has been
examined more closely (Roeder, 1967a; also fig. 3C, D).
When a moth is mounted in stationary flight (attached to
a support) and exposed to faint ultrasonic pulses from a
loudspeaker placed either to its right or its left, the moth
begins its attempt to turn away 45 milliseconds after the
beginning of stimulation. The experiments (fig. 3C, D)
show that it is able to choose the correct direction (right
or left) after receiving only the first pulse of the series, and
that it makes the change in flight direction by partially
folding its wings on the side of the body away from the
sound source.

THE ACOUSTIC SIGNAL AND THE EAR OF THE MOTH

These facts narrow the search for what takes place be-
tween the arrival of a stimulus and the change in flight
direction. Two other circumstances give additional en-
couragement to the search.

First, the cries made by a bat (fig. 4 A, B) can be dupli-
cated electronically (fig. 4C) with sufficient accuracy to
produce turning-away behavior. The artificial signal may

be said to have five different parameters or dimensions,
each of which can he \aried independently. It is possible,
therefore, to determine what aspects of the cries of a bat
will release and steer the evasive behavior of a moth. The
five parameters of the stimulus (fig. 4D) are: ( 1 ) the fre-
quency (pitch) of each sound pulse; (2) the amplitude
(intensity) of each pulse: (3) the duration of each pulse;
(4) the interval between pulses (repetition rate); and
i 5 ) the duration of the whole pulse train.

The present question is: How are these parameters trans-
lated or encoded by nerve-impulse patterns coming from
the ear of the moth and integrated by interneurons in its
central nervous system? The question may be put slightly
differently: Which of these parameters is significant in
determining what the moth finally does?

The second encouraging circumstance is the extreme
anatomical simplicity of the ear of a moth, which was
pointed out more than forty years ago (Eggers. 1925). A
noctuid moth has only two receptor cells in each ear, com-
pared with about fifty thousand in each ear of a human
being. Such a difference is a striking example of the parsi-
monious distribution of neurons in insect nervous systems
mentioned above. Practically, it simplifies the task of read-
ing out and assessing the total information reaching the
central nervous system of the moth via the channel that
connects it with the outside world. Electrodes can be placed
on the acoustic nerve, and the spike patterns delivered by
these two sense cells are readily interpreted under different
conditions of stimulation.

The details of the ear of a moth are shown in figure 5.
The bipolar sense cells (A, and A 2 ) are connected to the
eardrum by fine and complex organelles that transduce the
acoustic energy into a train of nerve impulses. The central
ends of A, and A 2 extend as two nerve fibers in the tym-

TYMPANIC

MEMBRA f IE

Fig. 5. Diagram of dorsal view of the right tympanic organ of a noctuid
moth. The tympanic membrane faces obliquely rearward and outward into the
constriction between thorax and abdomen. The scoloparium is a thin strand of
tissue attached to the inner surface of the tympanic membrane and suspended
in the air-filled sac by a ligament (top). The acoustic sense cells, Aj and A 2 ,
lie in the scoloparium. Their distal processes, extending toward the tympanic
membrane, transform sound energy into a series of nerve impulses, transmitted
to the central nervous system by the Ai and A» nerve fibers. The B fiber arises
from a non-acoustic sense cell, serving probably to register mechanical distor-
tions of the tympanic organ. Courtesy of Scientific American.

panic nerve connecting with the pterothoracic ganglion.
This nerve mass, which consists of the second and third
thoracic ganglia, is the site for the major neuronal trans-
actions concerned in bat avoidance.

NEURAL TRANSFORMATIONS AND TRANSACTIONS

The traffic of nerve impulses flowing from the tympanic
organ to the central nervous system is detected by an
electrode placed on the tympanic nerve. The sequence of
frames (fig. 6) shows how the spike patterns generated by
the more sensitive sense cell (A,) changes as the intensity

of a brief, ultrasonic pulse is increased by measured steps.
As the sound becomes louder, the spike pattern changes in
several respects: (a) more spikes are generated, that is.
a longer train is produced, although the duration of the
stimulus remains constant; (b) the spikes are more closely
spaced; (c) the latency or interval between the stimulus

Fig. 6. Spike responses (upper traces) recorded from an electrode on the
tympanic nerve of the moth, Xylena curvimacula, when pulses of 25 kilohertz
and 5 milliseconds in duration (middle traces) were directed at the ear. Sound
intensities are given in decibels above an arbitrary value (0) producing a minimal
acoustic response. Time marker (lower traces), 1000 cycles per second.

and the first spike of the series becomes less; and (d) at
lower intensities only the sense cell A, is stimulated, where-
as, at sound intensities ten times greater, it is joined by
responses of the A. sense cell (not shown in fig. 6).

Stated in another way, a hypothetical homunculus, sta-
tioned at the central termination of one tympanic nerve
in the thoracic ganglion, could determine intensity differ-

ences in the stimulus by four different criteria, not all of
them equally good. Criterion "a" might be ambiguous in
pulse duration, a longer pulse being confused with a louder
pulse. Criterion "b" would give a fairly accurate measure
of differences in pulse loudness. Criterion "c" would be
useful to the homunculus only if he could compare signals
coming from the right and left ears in response to the same
sound pulse. Criterion "d" would be a rough measure and
useful only in comparing very large differences in sound
intensity. Because we are concerned solely with the neu-
ronal mechanism of turning-away, which occurs at intensi-
ties capable of exciting only the A, sense cells, criterion
"d" can be neglected.

The same experiment, carried out with sound pulses of
different frequency (parameter 1), gives the same results
over a wide frequency range, roughly 15 to 100 kilohertz.
Thus, the moth appears to be tone deaf. The homunculus
could not measure parameter 1 from the spike signals
reaching him, although inspection of figure 6 shows that
the other four parameters of the stimulus are measured in
the spike pattern generated by the sense cells.

The next step is to find interneurons influenced by the
Aj signal and to determine in what ways they further trans-
form the spike pattern. A metallic microelectrode is low-
ered into the ganglion and used as an electrical probe. It
is moved about in search of the A, signal and of events
showing some causal relation to it.

From here on the trail becomes confused by a babel of
spike patterns, mostly of unknown origin and significance.
The A, pattern is easily recognized (fig. 7A). It reaches
the ganglion 3 to 5 milliseconds after sound reaches the
tympanic organ. Downstream from this point in the neu-
ronal mesh a number of interneurons have been encoun-
tered whose signals show various types of relation to the

A, response (Roeder, 1966b). I shall mention only three
of these, as they hint at ways in which the central nervous
system may convert stimulation into behavior.

The pulse-marker neuron is excited by a train of three

AV a WaVaWaWaVaWA

Fig. 7. Responses to stimulation of the tympanic organ recorded with micro-
electrode from sensory and interneurons in the pterothoracic ganglion of noctuid
moths (Caenurgina erechtea and Heliothis zea). A. Ai spikes recorded from
neuropile in response to a 5-millisecond ultrasonic pulse (middle trace); time
( lower trace ). 1000 cycles per second. B. Ai spikes (small downward deflections)
and single pulse-marker spike (large upward deflection) in response to 5-milli-
second sound pulse (lower trace). Time, 2 milliseconds per division. C. The
same, response to a longer (38-millisecond) sound pulse. Time, 5 milliseconds
per division. D. Single pulse-marker spike recurring in response to each of a
series of short ultrasonic pulses (lower trace) repeated 40 times a second.
E. Train-marker response. Spikes are indicated as dots on a raster that should
be read like consecutive lines on a printed page. Groups of larger dots are Ai
spikes, and indicate parameters 2-5 of the stimulus. Smaller dots are train-marker
spikes, recurring at a frequency independent of the pulse repetition rate through-
out the stimulation period. F. Change in the pattern of spikes in the motor nerve
supplying a muscle controlling extension of the forewing in response to stimula-
tion of the tympanic organ (middle trace). Motor response begins about 20
milliseconds after first sound pulse reaches the ear. Second sound pulse appears
to have no effect. Time, 100 cycles per second.

or four A, impulses coming from the ear on the same side
of the body (fig. 7B). The synaptic effects of the A, im-
pulses produce sufficient summation to trigger the pulse-
marker only if they are separated by intervals of 2
milliseconds or less. Typically, the response of the pulse-
marker is a single large spike, irrespective of the duration
of the stimulating sound pulse (parameter 3) and of the
resulting train of A spikes that impinges upon it. This
curious behavior of the pulse-marker (one spike per ultra-
sonic pulse irrespective of its duration) seems to depend
on a neuronal mechanism that requires 4 or 5 milliseconds
without synaptic bombardment by A, impulses in order
that the interneuron be "reset" to respond. This pause does
not occur when a long and moderately intense pulse reaches
the ear (fig. 7C). The pulse-marker will, however, generate
spikes up to 40 times per second if the long pulse is broken
into short pulses (fig. 7D). This behavior is interesting in
three respects.

First, the pulse-marker spike transmitted downstream
can be said to have discarded parameters 2 and 3 as
defined by the original ultrasonic stimulus. A homunculus
observing only the signal generated by one pulse-marker
could not judge differences either in the intensity or in the
duration of the original stimulus. He could still determine
pulse intervals (parameter 4) and the duration of the pulse
train (parameter 5).

Second, pulse-markers connected to the right and left
ears and sending their spikes into a mechanism that com-
pared relative times of arrival would be capable of steering
a moth in flight away from a distant sound source, because
the latency of the pulse-marker spike is long and variable,
depending as it does on the arrival of three or four A spikes
at sufficiently short intervals. Therefore the latency is
inversely related to intensity, and relative intensity (right
versus left ear) could be determined by marking whether

the right or the left pulse-marker fired first in response to
a given pulse. A neuronal mechanism making sueh a com-
parison has not yet been found.

Third, the behavior of the pulse-marker shows a striking
correlation with the behavior of flying moths exposed to
different ultrasonic pulse patterns (Roeder, 1964, 1967a).
Long, continuous tones produce only transitory turning-
away or none at all, whereas pulsed ultrasound causes a
sustained attempt to turn.

Among the neurons the signals of which have been
intercepted, two others seem relevant to the present account.
The first has been termed the train-marker neuron.

The train-marker neuron is inactive during silence, but
begins to discharge a train of spikes at an independent
frequency throughout the period in which a train of ultra-
sonic pulses reaches the ear (fig. 7E). The spike repetition
rate of the train-marker bears no relation to the pulse
repetition rate of the stimulus. Thus, the homunculus pro-
vided only with the train-marker signal would be able to
measure only the duration of a pulse train (parameter 5).
The other parameters of the stimulus would be lost to him.

Another interneuron, rarely encountered, appears to add
the A T signals coming from the right and left ears. It fires
twice as many spikes when both ears are stimulated as
when either ear alone is exposed to ultrasonic pulses.

TURNING-AWAY

These and other bits of information (Roeder, 1966b,
1967b) are insufficient for a definition of the neuronal
mechanism that is responsible for turning-away behavior.
The reason may be likened to the uncertainty principle in
physics — the deeper one searches for answers the greater
is the disturbance created by one's searching methods in the
beautifully poised living system. But one is heartened by

the hope that the biological obstacles are mainly technical
rather than theoretical as in the uncertainty principle
facing the physicists.

Many attempts have been made to approach the turning-
away response from the motor end, but the percentage of
success has been very small. In a few cases changes in the
pattern of motor impulses traveling to the wing-folding
muscles have been registered when the ear had been stimu-
lated with ultrasonic pulses (fig. 7E).

Incomplete though they are, the data presented give

TABLE 1

Parameters of the Ultrasonic Stimulus Present

in Various Signal Patterns

Signal
Pattern of

1 2

Frequency Amplitude

Duration

Pulse

Interval

5
Train
Length

Stimulus
Tympanic nerve
Pulse-marker
Train-marker
Turning-away behavior

X
X

X
X
X

X
X
X
X
X

" Not present at the pulse repetition rates in bat cries.

hints as to the kind of processing occurring in the central
nervous system. The original stimulus had five variable
parameters. The first of these, frequency (parameter 1),
is omitted from the tympanic nerve signal. Similar stages
in which other parameters are discarded are represented by
the pulse-marker (parameters 2 and 3) and the train-
marker (parameter 4) interneurons (table 1). It is as if
each parameter present in the original stimulus is a separate
key that permits admittance to a specific door but becomes
useless once the door in question has been passed.

At this point it seems worthwhile to compare the in-
formational content regarding the original stimulus that

is contained in the signal patterns registered at various
points in the nervous system of the moth with that con-
tained in its ultimate reaction — the turning-away behavior.
Such a comparison is summarized in table 1 . A train of
sound pulses reaching one ear at higher intensity causes a
steady, sustained attempt to turn away from the stimulus.
A continuous tone or a single pulse causes only a transitory
turning attempt. As the table shows, an observer of this
behavior could infer from it only the direction and the
pulse-train length of the stimulus; none of the other param-
eters of the original stimulus would be reflected in the
response. The neurophysiological experiments summarized
in table 1 suggest some of the steps in this elimination of
stimulus parameters as nerve signals propagate through
the nervous system and eventually shape the reaction of a
moth to a passing bat.

CONCLUSION

The mechanisms whereby nervous systems generate adap-
tive behavior have been regarded from three different
viewpoints. The first, which might be called the "center"
viewpoint, observes changes in behavior following relatively
massive surgical interference with the sense organs or with
parts of the central nervous system. It provides only a
broad picture of the functional topology of the nervous
system, and leads to concepts of "regions," or "centers,"
interacting with each other. The center viewpoint has been
of particular heuristic value in anlayzing insect nervous
systems because insect ganglia show a high degree of ana-
tomic separation, which is to some extent correlated with
function. For instance, it suggests that the insect brain
determines the "oneness," or "singularity," that is so uni-
versal in animal behavior. Such determination is accom-
plished under given conditions by the inhibition of all but

one of the action patterns organized by the segmental
ganglia.

At the present time, the center viewpoint seems un-
related to the much closer and more fine-grained viewpoint
of modern neurophysiology. Indeed, its conclusions do not
even require the postulation of neurons, nerve impulses,
or synapses. When regarded from the viewpoint of neuro-
physiology, the widely separated phylogenies of insect and
vertebrate nervous systems find a common base in the
behavior of single neurons. But when observed only from
the tip of a mircoelectrode differences between these two
groups of animals become mainly quantitative. There, the
second, or neuron, viewpoint also has its limitations.

The third viewpoint takes the findings of neurophysiol-
ogy for its basic assumptions. Given the intramural prop-
erties of neurons, it is concerned with neuron interaction,
with information transfer in neuron populations, and with
the way these and other functions could transpose a stimu-
lus pattern significant in the life of an animal into a re-
sponse promoting its survival.

At present, the neuron communities that can be com-
prehended from this viewpoint are small and simple, several
orders of magnitude simpler than those that are the concern
of the center viewpoint. Attempts to describe behavior in
terms of neuron populations are in their infancy. An infant
can handle, however, only simple toys, and hence I believe
that the simpler neuron communities — the nervous systems
of insects — have much to offer.

LITERATURE CITED

Adrian, F. D.

1930. The activity of the nervous system of the caterpillar. Jour. Physiol.,
vol. 70. pp. 34-35.

Eccles, J. C.

1953. The neurophysiologies! basis of mind. Oxford. Clarendon Press,

314 pp.
1964. The physiology of synapses. New York, Academic Press, 316 pp.

Eggers, F.

1925. Versuche liber das Gehor der Noctuiden. Zeitschr. f. Vergl. Physiol.,
vol. 2. pp. 197-314.

Griffin, D. R.

1958. Listening in the dark. New Haven, Yale University Press, 413 pp.

Hi ber. F.

1960. Untersuchungen iiber die Funktion des Zentralnervensystems und inbe-
sondere des Gehirnes bei der Fortbewegung und Lauterzeugung der
Grillen. Zeitschr. f. Vergl. Physiol., vol. 44, pp. 60-132.

1967. Central control of movements and behavior of invertebrates. In
Weirsma, C. A. G. (ed.). Invertebrate nervous systems. Chicago, Uni-
versity of Chicago Press, 370 pp.

Roeder, K. D.

1935. An experimental analysis of the sexual behavior of the praying mantis.

Biol. Bull., vol. 69, pp. 203-220.
1937. The control of tonus and locomotor activity in the praying mantis.

Jour. Exper. Zool., vol. 76, pp. 353-374.
1955. Spontaneous activity and behavior. Sci. Monthly, vol. 80, pp. 363-370.
1962. The behaviour of free flying moths in the presence of artificial ultra-
sonic pulses. Animal Behaviour, vol. 10, pp. 300-304.
1964. Aspects of the noctuid tympanic nerve response having significance in

the avoidance of bats. Jour. Insect Physiol., vol. 10. pp. 529-546.
1966a. Acoustic sensitivity of the noctuid tympanic organ and its range for the

cries of bats. Ibid., vol. 12, pp. 843-859.
1966b. Interneurons of the thoracic nerve cord activated by tympanic nerve

fibers in noctuid moths. Ibid., vol. 12, pp. 1227-1244.
1967a. Turning tendency of moths exposed to ultrasound while in stationary

flight. Ibid., vol. 13. pp. 890-923.
1967b. Nerve cells and insect behavior. Cambridge, Harvard University Press,

238 pp.

Roeder, K. D., L. Tozian, AND E. A. Weiant

1960. Endogenous nerve activity and behaviour in the mantis and cockroach.
Jour. Insect Physiol., vol. 4, pp. 45-62.

Rowell, C. H. F.

1965. The control of reflex responsiveness and the integration of behaviour.
In Treherne, J. E., and W. L. Beament (eds. ), The physiology of the
insect central nervous system. New York, Academic Press, 277 pp.

Wilson, D. M.

1961. The central nervous control of flight in a locust. Jour. Exp. Biol., vol.
38, pp. 471-490.

Wilson, D. M.

1967. An approach to the problem of control of ryhthmic behavior. In
Wiersma, C. A. G. (ed.), Invertebrate nervous systems. Chicago, Uni-
versity of Chicago Press, 370 pp.

tar

THIRTY-NINTH

JAMES ARTHUR LECTURE ON

THE EVOLUTION OF THE HUMAN BRAIN

1970

WHAT MAKES

MAN HUMAN

KARL H. ^PRIBRAM

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1971

THIRTY-NINTH

JAMES ARTHUR LECTURE ON

THE EVOLUTION OF THE HUMAN BRAIN

THI

THIRTY-NIN I H
JAMF.S ARTHUR LECTURE ON
EVOLUTION OF IMF HUMAN BRAIN

1970

WHAT MAKES MAN HUMAN

KARL H. PRIBRAM

Professor and Head

Neuropsychology Laboratories

Stanford University

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1971

LIBRARY

OF THE

JAMF.S ARTHUR LECTURES ON
THI EVOLUTION OF THE HUMAN BRAIN

Frederick Tilney, The Brain in Relation to Behavior; March 15, 1932

C. Judson Herrick, Brains <m Instruments of Biological Values; April 6, 1933

D. M. S. Watson, The Story of Fossil Brains from I'isli to Man; April 24, 1934

C. U. Ariens Kappers, Structural Principles in the Nervous System; The
Development of the Forehrain in Animals and Prehistoric Human Races;
April 25, 1935

Samuel T. Orton, The Language Area of the Human Brain and Some of its
Disorders; May 15, 1936

R. W. Gerard. Dynamic Neural Patterns; April 15, 1937

Franz Weidenreich, The Phylogenetic Development of the Hominid Brain and its
Connection with the Transformation of the Skull; May 5, 1938

G. Kingsley Noble, The Neural Basis of Social Behavior of Vertebrates; May
11, 1939

John F. Fulton, A Functional Approach to the Evolution of the Primate Brain;
May 2, 1940

Frank A. Beach, Central Nervous Mechanisms Involved in the Reproductive
Behavior of Vertebrates; May 8, 1941

George Pinkley, A History of the Human Brain; May 14, 1942

James W. Papez, Ancient Landmarks of the Human Brain and Their Origin;
May 27, 1943

James Howard McGregor, The Brain of Primates; May 11, 1944

K. S. Lashley, Neural Correlates of Intellect; April 30, 1945

Warren S. McCulloch, Finality and Form in Nervous Activity; May 2, 1946

S. R. Detwiler, Structure-Function Correlations in the Developing Nervous
System as Studied by Experimental Methods; May 8, 1947

Tilly Edinger, The Evolution of the Brain; May 20, 1948

Donald O. Hebb, Evolution of Thought and Emotion; April 20, 1949

Ward Campbell Halstead, Brain and Intelligence; April 26, 1950

Harry F. Harlow, The Brain and Learned Behavior; May 10, 1951

Clinton N. Woolsey, Sensory and Motor Systems of the Cerebral Cortex; May
7, 1952

Alfred S. Romer. Brain Evolution in the Light of Vertebrate History; May 21,
1953

Horace W. Magoun, Regulatory Functions of the Brain Stem; May 5, 1954

*Fred A. Mettler. Culture and the Structural Evolution of the Neural System;
April 21, 1955

*Pinckney J. Harman, Paleoneurologic, Neoneurologic, unci Ontogenetic Aspects
of Brain Phylogeny; April 26, 1956

* Davenport Hooker, Evidence of Prenatal Function of the Central Nervous

System in Man; April 25, 1957

* David P. C. Lloyd, The Discrete and the Diffuse in Nervous Action; May 8,

1958

"Charles R. Noback, The Heritage of the Human Brain; May 6, 1959

*Emst Scharrer, Brain Function and the Evolution of Cerebral Vascularization;
May 26, 1960

Paul I. Yakovlev, Brain, Body and Behavior. Stereodynamic Organization of the
Brain and of the Motility-Experience in Man Envisaged as a Biological
Action System; May 16. 1961

H. K. Hartline, Principles of Neural Interaction in the Retina; May 29, 1962

Harry Grundfest, Specialization and Evolution of Bioelectric Activity; May 28,
1963

"Roger W. Sperry, Problems Outstanding in the Evolution of Brain Function;
June 3, 1964

"Jose M. R. Delgado, Evolution of Physical Control of the Brain; May 6, 1965
Seymour S. Kety, Adaptive Functions and the Biochemistry of the Brain; May
19, 1966

Dominick P. Purpura, Ontogenesis of Neuronal Organizations in the Mammalian
Brain; May 25, 1967

"Kenneth D. Roeder, Three Views of the Nervous System; April 2, 1968

tPhillip V. Tobias, Some Aspects of the Fossil Evidence on the Evolution of the
Hominid Brain; April 2, 1969

"Karl H. Pribram, What Makes Man Human; April 23, 1970

"Published versions of these lectures can be obtained from The American
Museum of Natural History, Central Park West at 79th St., New York,
N. Y. 10024.

tTo be published for The American Museum of Natural History by the
Columbia University Press.

The hippopotamus may well regard man, with his physical weakness,
emotional unpredictability, and mental confusion as a freak. . . .

(Heschel, 1965, p. 23)

CONTENTS

Introduction 1

What A Code Is 4

Brain Function in Awareness

The Motor Mechanism and Acts 12

Signs and Symbols: Association or Differentiation? 18

Propositions and Reasoning: Using Signs Symbolically and

Symbols Significantly 26

Conclusion 31

References Cited 33

WHAT MAKES MAN HUMAN

Introduction

/\ //// MIDDLl li. is THINKERS WERE
TRYING l<> DISCOVER PROOFS FOR i m

I Ms i / \< l <>l (.on TOD n ll /■ SEEM l <>
LOOK FOR PROOF I OR I III t.XISIIXCI: Ol

M '\ (III s< in i . i< lt , r „ p, 26)

What makes man human is his brain. This brain is
obviously different from those of nonhuman primates. It is
larger (Jerison. 1961), shows hemispheric dominance and
specialization (Mountcastle, 1962), and is cytoarchitec-
turally somewhat more generalized (Bailey and von Bonin,
1951; Lashley and Clark. 1945). But are these the
essential characteristics that determine the humanness of
man? This paper cannot give an answer to this question for
the answer is not known. But the problem can be stated
more specifically, alternatives spelled out on the basis of
available research results, and directions given for further
inquiry.

My theme will be that the human brain is so constructed
that man, and only man, feels the thrust to make meaningful
all his experiences and encounters. Development of this
theme demands an analysis of the brain mechanisms that
make meaning — and an attempt to define biologically the
process of meaning. In this pursuit of meaning a fascinating
variety of topics comes into focus: the coding and recoding
operations of the brain; how it engenders and processes
information and redundancy; and, how it makes possible
signs and symbols and propositional utterances. Of these,
current research results indicate that only in the making of
propositions is man unique — so here perhaps are to be
found the keynotes that compose the theme.

My concern with meaning originated in an attempt to
formulate what ails the current educational process
(Pribram, 1964, 1969a). Education entails communication
between generations. As such, educational institutions have
been set up to transmit information. Our schools have
rightly been occupied with problems of information storage
and retrieval: what ought to be taught in what period of
time and how it is to be demonstrably retrieved.

But, it seems this is not enough. From those whom we
try to educate we hear rumblings and even shouts of
discontent — discontent which arises at least in part from
our failure to meet an educational need. What might this
be? Is the mere acquisition of information insufficient? May
the accumulation of information even be a cause of the
problem? Is it not imperative to attempt to impart some-
thing additional, something which makes information
meaningful?

Information measurement theory provides an interesting
starting point for inquiry into this question. In an organism
endowed with memory the acquisition of information can,
on occasion, actually lead to an increase in uncertainty.
Take, for instance, a family. The wife is at home, her
husband away on a trip, and two children are in college.
Her husband informs her that he will call on Thursday, her
birthday. Letters from the children give the additional
information that they also will call. When the phone rings
the wife experiences an amount of uncertainty equivalent
to the amount of information she was given initially. She
can reduce her uncertainty by obtaining more "informa-
tion": asking who's calling. But note that though at the
moment of the call the answer to her question provides
information, when the extended time period over which
the entire episode has transpired is considered, the answer

is a repetition of one of the earlier messages. Thus, over

time, uncertainty is countered, not by something novel, not
b\ information, but by redundancy, i.e., by repetition.

My thesis will be that meaning — the gerund of an old
English word for intend, give purpose to — is made possible
by repetition. Let me spell out this thesis, first in general,
then in brain terms. Repetition comes in many forms. Some
forms, some patterns of repetition, are more meaningful
than others. Patterns of repetition are called codes. Codes
are constructed for a useful purpose. When an organism
is uncertain he has two alternative strategies to follow: one,
he can reduce uncertainty by seeking real novelty, i.e.,
information. This, as already noted, will often bring only
temporary relief because of man's mnemonic capacity. The
other strategy is to reduce uncertainty by coding — by
enhancing redundancy, repeating the familiar. This carries
the penalty of boredom unless the patterns of repetition are
varied. Varying a code turns out to be a remarkably
powerful instrument for effectively reducing uncertainty
because it permits using information in unexpected ways.

From my own research I have concluded that one of the
most pervasive — perhaps the most pervasive — of the opera-
tions of the brain is. when the need is felt, to actively revise
the patterns of redundancy in which information is encoded
(Pribram. 1969b). There are several levels of these en-
coding operations, each useful in its own way. Let me first
say something about what a code is and then describe the
types of codes constructed by the brain.

What A Code Is

WOXDER, OR RADICAL AMAZEMENT, IS A
WAY OF GO/.YG BEYOND WHAT IS GIVEN IX
THIXG AND THOUGHT. REFUSING TO TAKE
AXYTHIXG FOR GRANTED, TO REGARD ANY-
THING AS FINAL." (HESCHEL, 1965, pp. 78-79)

Not so long ago my laboratory came into the proud
possession of a computer. Very quickly we learned the fun
of communicating with this mechanical mentor. Our first
encounter involved twelve rather mysterious switches which
had to be set up (U) or down(D) in a sequence of patterns,
each pattern to be deposited in the computer memory before
resetting the switches. Twenty such instructions or patterns
constituted what is called the "bootstrap" program. Only
after this had been entered could we "talk" to the computer
— and it to us — via an attached teletype. For example:
DUUUUUDDDDUD
DDDDDDDDUUDD
DUUUDDUDUUUD
DUDDUDDDUUDU
DDUDDDDUUUUUandsoon.
Bootstrapping is not necessarily an occasional occurrence.
Whenever a fairly serious mistake is made — and mistakes
were made often at the beginning — the computer's control
operations are disrupted and we must start anew by
bootstrapping.

Imagine setting a dozen switches twenty times and
repeating the process from the beginning every time an
error is committed. Imagine our annoyance when the
bootstrap didn't work because perhaps on the nineteenth
instruction an error was made in setting the eighth switch.
Obviously, this was no way to proceed.

Computer programmers had early faced this problem and

solved it simply. Conceptually, the twelve switches are

divided into four triads and each combination of up or

down within each triad is given an Arabic numeral. Thus,

D D D became

became

U U D became 6

U U U became 7
Conceptually, switching the first toggle on the right becomes
a 1 , the next left becomes a 2, the next after this a 4, and
the next an 8. If more than a triad of switches had been
necessary, if, for instance, our computer had come with
sixteen switches, we should have conceptually divided the
array into quads. Thus the bootstrapping program now
consisted of a sequence of twenty patterns of four Arabic
numerals, such as:

3 7 2 2

3 4 5 6

2 2 15

10 3 7 etc.
and we were surprised at how quickly those who boot-
strapped repeatedly, actually came to know the program
by heart. Certainly fewer errors were made in depositing
the necessary configurations — the entire process was speed-
ed and became, in most cases, rapidly routine and habitual.
Once the computer is bootstrapped it can be talked to
via a teletype in simple alphabetical terms, for example,
JMP for jump, CLA for clear the accumulator, TAD for

add, etc. But each of these mnemonic devices merely stands
for a configuration of switches. In fact, in the computer
handbook the arrangement for each mnemonic is given in
Arabic notation: e.g. CLA = 7200. This in turn is easily
translated into UUUDUDDDDDDD should we be
forced to set the switches on the computer by hand because
the teletype has gone out of commission.

In the first instance, then, programming is found to be
the art of devising codes, codes that when hierarchically
organized facilitate learning, remembering and reasoning.
The power of the coding process is not to be underestimated.
Should you doubt this, try next month to check your bank
statement against your record of expenditures and do it all
using Roman rather than Arabic numerals. Can you imagine
working out our national budget in the Roman system?

Next let me turn to an analysis of the classes of codes
engendered by the brain. These must account for the exist-
ence of subjective states such as perceptions and feelings;
for the achievement of acts in the organism's environment;
for the construction of signs and symbols by which organ-
isms communicate with each other; and for the composition
of propositions, the tools with which man reasons and has
fashioned his culture. Research on the brain mechanisms
relevant to each of these classes has in recent years yielded
some fascinating surprises (Pribram, 1971). Let me share
some of these surprises with you in the search for meaning
even if at times the connection between brain, behavior
and meaning will appear to be remote. My route is a
deliberate one, however, because for me: "Knowing [about
meaning has not been] due to coming upon something,
naming and explaining it. Knowing has been due to some-
thing forcing itself on [me]."' (Heschel, 1965. p. 109).

Brain Function in Awareness

//// EXPERIENCl 01 I MEANING IS AN

I \/7 nil \< / 01 i 1 1 ii INVOLVEMEN1 ...
\<>l i\ EXPERIENCl a I PR/1 ill REFER
I \< l 01 VI i\l\(.. i:i I \u IRING I DIMl S
sio\ 0P1 \ l(> H I ill \l l\ HI l\(.s

(HESCHEL, 196$, p. 79)

During the past decade a series of studies initiated by
Kamiya (1968) has shown that people can discriminate
their brain states. These studies use electrical signals to
indicate brain function and recordable behaviors as meas-
ures of psychological state. A subject readily acquires the
ability to discriminate the occasions when his brain is giving
off alpha rhythms from those when his brain's electrical
activity is desynchronized. An interesting incidental finding
in these studies has been the fact that when Zen and Yoga
procedures accomplish their aims, subjects can attain the
alpha brain rhythm state at will. Kamiya's training
procedures can and are being used as a short cut to Nirvana.

More specific are some recent experiments of Libet
(1966) that have explored a well-known phenomenon.
Since the demonstrations in the late 1 800's by Fritsch and
Hitzig (1870) that electrical stimulation of parts of man's
brain results in movement, neurosurgeons have explored its
entire surface to determine what reactions such stimulations
will produce in their patients. For instance, Foerster (1936)
mapped regions in the postcentral gyrus which give rise to
awareness of one or another part of the body. Thus sensa-
tions of tingling, of positioning, etc. can be produced in the
absence of any observable changes in the body part experi-
enced by the patient. Libet has shown that the awareness
produced by stimulation is not immediate: a minimum of a
half second and sometimes a period as long as five seconds

elapses before the patient experiences anything. It appears
that the electrical stimulation must set up some state in the
brain tissue and only when that state has been attained does
the patient experience.

What do we know about the organization of these brain
states apparently so necessary to awareness? They display
some curious properties. One would expect that when the
brain rhythms which are correlated with the subject's
report are disrupted, the behavioral functions would also
be interfered with. This is not the case. Focal epileptic
discharge in the postcentral gyrus (Stamm and Warren,
1961 ) and elsewhere, unless it becomes pervasive and takes
over the function of a large part of the brain, does not
seriously disrupt awareness. I have densely scattered epi-
leptic lesions in various areas of the nonhuman primate
brain in a series of carefully carried out experiments and
found that despite the electrical disturbance produced,
problem-solving ability remains unimpaired provided the
ability had been acquired before electrical seizure discharge
began (Kraft, Obrist and Pribram, 1960; Stamm and
Pribram. 1960; Stamm and Pribram, 1961 ). (The acquisi-
tion of appropriate performances after the discharges be-
come established is, however, slowed approximately
fivefold. )

In short, the brain state necessary to awareness appears
to be resistant to being disrupted by local damage provided
this damage is not overly extensive. An estimate of the
limits on the extent to which disruption can take place
without undue influence on the state comes from experi-
ments involving brain tissue removals. Some 85% (or in
some experiments even more, Galambos, Norton and
Frommer, 1967; Chow, 1970) of a neural system can be
made ineffective without seriously impairing the perform-

ances dependent on that system ( Lashley. 1950). What
sort oi state is it that ean function effectively when only
10 or 15 of it remains and all of what remains need not
he concentrated in one location?

The answer is that the effective units of the state must
be distributed across the tissue involved. Each unit or small
cluster of units must be capable of performing in lieu of the
whole. Until very recently it was difficult to conceive of
such a mechanism.

But just as information processing by computer is an
aid in conceptualizing the way in which coding operations
are hierarchically constructed, so another engineering
domain helps us to understand the problem of the "dis-
tributed" state. This domain is called optical information
processing (van Heerden. 1968) because optical systems
work this way; or. holography, because each part of a
recorded state can stand in for the whole (Leith and
Upatnicks, 1965).

The essential characteristic of a holographic state is
the encoding of the relation among recurrences of neigh-
boring activities. This is known technically as a spatial
phase relationship. In optics, ordinary pictures encode only
the intensity of illumination at any location; a hologram
encodes spatial phase in addition.

Holograms have many properties of interest to the brain
scientist. Foremost of these is the fact that information is
distributed in the holographic record. Thus one can take
a small part of the hologram and reconstruct from it an
image in most respects the same as that reconstructed from
the whole record. Second, a great deal of information can
be stored in one hologram. Several major companies (IBM,
RCA ) have been able to encode well over a million bits in
a square centimeter. Third, an entire image can be recon-

structed from a hologram when illumination is reflected
from one feature or part of the scene originally recorded.
This is the property of associative recall.

Holograms were first constructed mathematically by
Dennis Gabor (1949, 1951) and crude reproductions were
achieved. Later they were improved immensely by illu-
minating the object with a laser beam. Because of the
similarity of properties of the optical hologram and the

MONOCHROMATIC
POINT LIGHT SOURCE

OBJECT

FOCAL PLANE
WITH FILTER

IMAGE

Fig. 1. The information to be stored is originally present on a transparent
slide in the object plane O. It is illuminated by parallel light from a coherent
light source L, like a laser beam. Consequently, in the image plane I one will
see an image of the transparent object, faithful within the limitations of the
optical system. We now expose a photographic plate, not in I, but in the focal
plane F, to the light diffracted by the object. This plate, after exposure, is de-
veloped and a positive is made of it, which is put back in F. This filter, which has
a transmission in each point proportional to the original light intensity, is called
a hologram.

facts about the brain reviewed in the passages above, I
have suggested that one important encoding process in the
brain follows the mathematical rules of holography
(Pribram, 1966). My laboratory is now working on the
problem of just how the hologram is realized in neural tissue
(Pribram, 1969a).

The neural hologram is a state in which information is
encoded in such a way that images can be constructed.

Although images are evanescent, they occur. Although they
cannot be directly communicated, they exist. At least three
types of images can be discerned subjectively, however, and
for each a separate neural system has been identified. Images
constructed by the operations of the classical sensory sys-
tems refer to events external to the organism (Pribram,
1966); images constructed by the operations of the limbic
forebrain monitor the world within (Pribram, 1967a;
Pribram, 1970); and, images constructed by the brain's
motor mechanisms structure the achievements an organism
aims to accomplish (Pribram, et al., 1955-56; Pribram,
1971). I want now to take a look at these motor mech-
anisms, for without them behavior could not occur and
we could never make our images meaningful.

The Motor Mechanism and Acts

•THE DEED IS THE DISTILLATION OF THE
SELF." (HESCHEL, jg^, p. g4)

Neuroscientists have engaged in a century-long con-
troversy regarding the functions of the motor cortex of the
brain. The view common to all protagonists has been that
this tissue serves much as does a keyboard upon which the
remainder of the brain — or the mind — constructs the
melodies to be executed by muscles as behavior (Sherring-
ton, 1906). What has been controversial is the nature of
the keyboard. Does it encode, i.e. contain a representation
of, individual muscles or even parts ( Woolsey, Chang and
Bard, 1947; Bucy, 1949); or, does the keyboard encode
movements, spatial and temporal combinations of muscle
contractions, much as do the more complex controls of an
organ which encode chords, timbre, etc. (Walshe, 1948;
Lashley, 1921)?

Some years ago I set out to see for myself where I stood
in this controversy. I repeated some of the classical experi-
ments and performed others. The results were surprising
and I was unable to understand them fully until very
recently when additional data from other laboratories
became available.

The first surprise came with the discovery that sensory
nerves from both skin and muscle send signals to the motor
cortex by pathways no more circuitous than those by which
such signals reach sensory cortex (Malis, Pribram and
Kruger, 1953). If the motor cortex were indeed the final
common path for cerebral activity, a funnel, what business
has it to be informed so directly from the periphery? The

problem was compounded by a scries of reports of experi-
ments analyzing the organization of peripheral motor con-
trol which appeared about this time (Granit, 1955; Granit
and Kellerth. 1967; Kuffler and Hunt. 1952). The results
of these experiments showed that one-third of the fibers
leaving the spinal cord destined for muscle end in muscle
receptors and have, under the experimental conditions, no
immediate influence on muscle contraction. What happens
when these fibers (called the y system because they are the
smallest in diameter) are stimulated electrically is that a

A B

Fig. 2. A. Cortical response evoked by stimulation of superficial peroneal nerve.
Upper trace in the postcentral "sensory" cortex; lower trace in the precentral
"motor" cortex. Time: 10 msec. B. Same as A except that stimulus was applied
to posterior tibial nerve. Note that the response in the "motor" cortex is prac-
tically identical to that in the "sensory" area.

Fig. 3. These responses were obtained
on sciatic stimulation after complete
resection of cerebellum plus additional
resection of cortex of both postcentral
gyri. Upper trace, postcentral exposed
decorticated white matter: lower trace,
precentral cortex. Time: 2 and 10 msec.
This indicates that the responses shown
in Fig. 2 do not traverse the sensory
cortex or the cerebellum on the way
to the "motor" cortex.

change is produced in the signals going to the spinal cord
from the muscle receptors. Until these experiments were
reported it had been thought that the signals from the
muscle receptors accurately reflected the states of contrac-
tion or relaxation of the muscles. Now it became necessary
to take into account the fact that messages from the central
nervous system could influence the muscle receptors in-
dependent of any changes produced in the muscle.

The results of both these sets of experiments spelled the
end to a simple stimulus-response model of how the nervous
system controls behavior (Miller, Galanter and Pribram,
1960). At the periphery the reflex arc became an untenable
fiction; at the cortex the keyboard had to give way to some
more sophisticated conception.

The second surprise regarding the motor mechanism
came with the discovery that I could remove huge amounts
of motor cortex with very little impairment of muscle
function (Pribram, et al., 1955-56). Neither individual
muscle contractions nor any particular movements were
seriously altered by the surgery. Yet something was amiss.
Certain tasks were performed with less skill despite the fact
that slow motion cinematography showed the movements
necessary to perform the task were executed without flaw in
other situations. My interpretation of this finding was that
behavioral acts, not muscles or movements, were encoded
in the motor cortex. An act was defined as an achievement
in the environment that could be accomplished by a variety
of movements which became equivalent with respect to the
achievement. Thus a problem box could be opened by use
of a right or left hand; amputees have learned to write with
their toes. Encoded in the motor cortex are the determinants
of problem solution and of writing — not the particular
movements involved in the performance.

What I could not fathom at the time was how the

determinants of an act could be encoded. Two experiments
have recent]) helped to clarify my perplexity. One was
performed by Bernstein ( 1967) in the Soviet Union. Bern-
stein photographed people clad in black leotards carrying
out preassigned tasks against black backgrounds. Patches
of white were attached to the leotards at the locations of
major joints. Examples of the tasks are hammering a nail
and running over rough terrain. Cinematography showed
only the white patches, of course. These described a running
wave form which could be analyzed mathematically. From
his analysis Bernstein could predict within 2 mm. where the
next movement in the action would terminate — where the
hammer blow would fall, what level the footsteps would
seek. It became obvious that if Bernstein could make such
a calculation, the motor cortex could also do it. Interest-
ingly, the equations Bernstein used were the temporal
equivalent of those which describe the hologram.

The second experiment gives a clue as to which deter-
minants of acts are encoded. Evarts ( 1967) impaled cells
in the motor cortex of monkeys with fine electrodes and
recorded the activity of these cells while the monkey pushed
a lever. Different weights were attached to the lever so
that greater or lesser force had to be exerted by the monkey
in order to accomplish the task. Evarts. to his surprise,
found that the activity of the cortical neurons from which
he was recording varied not as a function of the length or
stretch of the muscles used to push the lever but as a
function of the force needed to perform the task. Apparently
what is encoded in the motor cortex is a representation of
the field of forces describing the conditions necessary to
achieve an action.

Now the earlier experimental results began to make sense.
The motor mechanism resembles a set of thermostats rather
than a keyboard (Merton, 1953). At the periphery the
receptors are subject to a dual influence: they are sensitive
to muscle tension, which reflects the force exerted on the
muscle, and they are sensitive to signals from the central
nervous system by way of the y fibers. This is much like the
sensitivity of the thermocouple in a thermostat which is com-
posed of two pieces of metal separated when cool but which
make contact with each other by expanding when warmed.
In addition to the sensitivity to temperature change the size
of the gap between the pieces of metal can be varied by the
little wheel at the top of the thermostat — i.e. the device can
be set to be more or less sensitive to heat. There is by now
a large body of evidence that the y motor system works by
setting the muscle receptor's sensitivity to changes in muscle
tension (Mettler, 1967). There is also a great deal of evi-
dence that much of the brain's control over muscle function
is performed by making changes in set, in biasing the y sys-
tem, and not in making muscles move directly. Note that the
setting device of the thermostat is calibrated for tempera-
ture, that it has encoded on it the information necessary
to control the activity of the furnace to reach the goal set
for it and that this goal can be met over a wide range of
changes in the temperature of the environment. Note also
that the furnace need not display any fixed rhythm of on
and off — this rhythm will vary with the environmental exi-
gencies. In the same manner, the brains motor mechanism
can encode the set points, the information necessary to
achieve certain acts. The brain need not keep track of the
rhythms of contraction and relaxation of individual muscles
necessary to achieve an act any more than the thermostat
needs to keep track of the turnings on and off of the furnace.

The encoding problem is immensely simplified — only end
states need to be specified. As already noted these can he
computed h\ extrapolation from holographic -like equations
that summarize the sequence ( repetitions ) of forces ( muscle
tension states) exerted.

This is the manner in which the brain achieves acts. But
we are not yet arrived at meaning. Acts can be stereotyped,
routine. They can be made necessary by environmental
change, necessary merely to maintain the organism's equi-
librium in the face of such changes. No, there is more to
meaning than just action, as there is more to meaning than
just imaging. Meaning is derived when acts intend (from the
Latin intendere, to stretch toward), that is, reach out to.
thus impaling otherwise evanescent images and keeping
them from slipping away. The brain makes this possible
by constructing signs and symbols.

Signs and Symbols: Association or Differentiation?

"KNOWLEDGE IS FOSTERED BY CURIOSITY;
WISDOM IS FOSTERED BY AWE. AWE PRECEDES
FAITH; IT IS THE ROOT OF FAITH.''

(HESCHEL, 1965, p- 89)

Much of my own research on nonhuman primates has
been devoted to the problem of how the brain makes pos-
sible signs and symbols. For many years I questioned
whether, in fact, nonhuman primates could construct signs
and symbols but my doubts have now been resolved by work
with two chimpanzees, one studied by the Gardners ( 1969)
at the University of Nevada and one by Premack (1970) at
the University of California at Santa Barbara. The Nevada
chimpanzee named Washoe ( after the county in which Reno
is located) has been taught to communicate using a sign
language devised for the deaf and dumb. Earlier attempts
to set up a rich communicative system between chimpanzee
and man had failed. The Gardners felt that this failure was
due to the limitations of the chimpanzee vocal apparatus
and therefore decided to use a gestural system instead. The
system chosen, American Sign Language, has the added
feature that it is a relatively iconic rather than a phonetic
system, thus much less complex in its structure than is
human speech.

Washoe has learned to use approximately 150 signs. She
can string two or three signs together but not in any regu-
larly predictable order. Comparison with deaf human chil-
dren of comparable age shows marked differences in the
way in which gestural signs are used — but more of this later.
The point here is that sign making is possible for the non-
human primate.

The Santa Barbara chimpanzee, Sarah, is being trained
by an entirely different method to an entirely different pur-
pose. Premack has taken operant conditioning methods and
applied them to determine just how complex a system of
tokens can be used to guide Sarah's behavior. Experiments
performed in the 1930's had already shown that chim-
panzees will work for tokens — in fact a chimpomat had been
constructed for use with poker chips. The chimpomat was
an outgrowth of the delayed response task, the indirect form
of which uses a temporary token to indicate where a piece
of food (a reinforcer) is to be found subsequently. The de-
layed response task had been devised to determine whether
animals and children could bridge a temporal gap between a
momentary occurrence and a later response contingent on
that occurrence. The bridge, which animals and children
can construct, has been variously conceptualized in terms of
"ideas." "memory traces," "short term memory organiza-
tion." etc. Premack's chimpanzee has demonstrated that
behavior dependent on tokens is not only possible but that
hierarchical organizations of tokens can be responded to
appropriately.

In all of these experiments the crux of the problem is that
the token does not call forth a uniform response. Depending
on the situation, that is, the context in which the token
appears, the token must be apprehended, carried to another
location, inserted into a machine or given to someone, traded
for another token or traded in for a reward. Or, as in the
original delayed response situation, the token stands for a
reward which is to appear in one location at one time, an-
other location at another time.

I shall use the term "symbols" to describe these context
dependent types of tokens to differentiate them from "signs"
which refer to events independent of the context in which

they appear. ( This distinction is consonant with that made
by Chomsky [1963]. "Formal Properties of Grammars,"
and is used here to indicate that the primordia of the rules
that govern human language are rooted in what are here
called "significant'* and '"symbolic" processes.) There is now
a large body of evidence to show that the cortex lying be-
tween the classical sensory projection areas in the posterior
part of the brain is involved in behavior dependent on dis-
criminating signs and that the frontal cortex lying anterior
to the motor areas is involved in performances dependent
on symbolic processes.

The surprise came when experiments were devised to
show how these parts of the brain worked in determining
sign and symbol. The ordinary view is that progressively
more complex features are extracted or abstracted from in-
formation relayed to the projection areas: the simpler ex-
tractions occur in the projection areas per se, more complex
abstractions demand relays beyond this primary cortex to
adjacent stations where associations with information from
additional sources (e.g. the primary projection areas) are
made available (Hubel and Wiesel, 1965). Unfortunately
for this view there is a good deal of experimental evidence
against it.

Most direct is the fact that if progressive cortico-cortical
relays are involved in the ability to utilize signs and symbols,
then removals of these relays should impair the ability. This
is not the case. The posterior and frontal cortices specifi-
cally concerned in sign discrimination and in delayed re-
sponse lie some distance from the primary sensory and
motor areas. Complete removal of the tissue that separates
the primary areas from those involved in discrimination and
delayed response does not permanently impair the perform-
ance of these tasks (Lashley, 1950; Chow, 1952; Pribram,

Fig. 4. Diagrammatic reconstruction of the brain after an essentially complete
lesion of the peristriate cortex. Representative cross sections are shown by num-
ber indicating placement on brain diagram. The monkey from whom this brain
was taken retained a visual discrimination habit perfectly.

Spinelli and Reitz, 1969). Ergo, cortico-cortical "abstrac-
tive" relays cannot be the mechanism at issue.

Two possibilities remain to explain the involvement of
those cortical areas remote from the primary projection
zones in discrimination and delayed response behavior. In-
formation may reach these areas by routes independent of
those that serve the primary projection cortex. This possi-
bility is being actively explored in several laboratories. In
the rhesus monkey, however, there is already evidence that
these independent routes do not play the desired role: de-
struction of the pathways does not lead to a deficit in the per-
formance of discriminations or delayed response (Chow,
1954;Mishkin, 1969).

The third possibility is one that I have been seriously ex-
ploring for the past decade and a half (Pribram, 1958a).
This alternative holds that sign and symbol are constructed
by a mechanism that originates in the cortex and operates
on the classical projection systems in some subcortical loca-
tion. Thus the effects of the functioning of the cortex in-
volved in signing and symbolizing are conceived to be trans-
mitted downstream to a locus where they can preprocess
signals projected to the primary sensory and motor cortex.

A good deal of evidence has accrued to this third alterna-
tive. Perhaps most important is the fact that a large portion
of the pathway relays within the basal ganglia, motor struc-
tures of the motor mechanism of the brain (Reitz and Pri-
bram, 1969). Sign and symbol manipulation thus involves
the same brain structures that are used by the organism in
the construction of acts. The suggestion that derives from
these anatomical facts is that signifying and symbolizing
are acts, albeit acts of a special sort.

There is, of course, a difference in the neuroanatomy in-
volved in signifying and that involved in symbolizing. This

Putamen
A18

Fig. 5. Responses evoked by stimulation of the part of the temporal lobe in-
volved in vision. Note tracts passing through the putamen, one of the major
motor structures in the brain. Horizontal marks indicate the location of the
tip of the recording electrode from which the response was photographed.

difference, as well as the behavioral analysis or the tasks
involved, tells a good deal of what these behavioral processes
are all about. The pathways for signifying influence the pri-
mary sensory systems. Connections have been traced by elec-
trophysiological techniques as far peripheral as the retina
(Spinelli, Pribram and Weingarten, 1965; Spinelli and Pri-
bram, 1966) and the cochlear nucleus (Dewson, Nobel and
Pribram, 1 966), for instance. The connections important to
the symbolic process have not as yet been determined as
fully, but a good deal of the evidence points to involvement
with the limbic systems structures on the innermost bound-
ary of the forebrain (Pribram, 1958b) .

This connection between limbic and frontal lobe func-
tion demands a word or two. Removal of tissue in these
systems does not impair sign discrimination but does impair
performance on such tasks as delayed alternation (Pribram,
et al., 1952; Pribram, et al., 1966; Pribram, Wilson and
Connors, 1962), discrimination reversal (Pribram, Douglas
and Pribram, 1969), shuttle-box-avoidance (Pribram and
Weiskrantz, 1957) and approach-avoidance, commonly
called "passive" avoidance (McCleary, 1961). In all of
these tasks some conflict in response tendencies, conflict
among sets, is at issue. The appropriate response is context
(i.e. state) dependent and the context is varied as part of
the problem presented to the organism. Thus a set of con-
texts must become internalized (i.e. become brain states)
before the appropriate response can be made. Building sets
of contexts depends on a memory mechanism that embodies
self referral, rehearsal or, technically speaking, the operation
of sets of recursive functions. (The formal properties of
memory systems of this type have been described fully by
Quillian, 1967.) The closed loop connectivity of the limbic
systems has always been its anatomical hallmark and makes

an ideal candidate as a mechanism tor context dependency
( Pribram. 1961 ; Pribram and Kruger, 1954).

As an aside, it is worth noting that much social-emotional
behavior is to a very great extent context dependent. This
suggests that the importance of the limbic formations in
emotional behavior stems not only from anatomical connec-
tivity with hypothalamic and mesencephalic structures but
also from its closed loop, self-referring circuitry. It remains
to be shown ( although some preliminary evidence is at hand
| Fox, et al., 1967; Pribram. 1967b]) that the anterior
frontal cortex functions in a corticofugal relation to limbic
system signals much as the posterior cortex functions to
preprocess sensory signals.

Thus signs and symbols are made by the brain's motor
mechanism operating on two classes of images — in the case
of signs those that encode sensory signals and in the case
of symbols those that monitor various states of the central
nervous system. Signs are codes invariant in their reference
to events imaged — their meaning is context free. The mean-
ing of symbols, on the other hand, is context dependent and
varies with the momentary state induced in the brain by the
stimulation. Both signs and symbols convey meaning, make
possible a temporal extension of otherwise momentary
occurrences.

Man shares the meaning conveyed by sign and symbol
with nonhuman animals. This form of meaning, though per-
haps more highly developed in man than in other animals, is
not what makes him peculiarly human. Our search for man's
unique thrust to make all his experiences and encounters
meaningful needs to proceed to yet another level of com-
plexity of encoding: only man makes propositions and rea-
sons with them.

Propositions and Reasoning: Using Signs Symbolically and
Symbols Significantly

"MAN MAY, INDEED, BE CHARACTERIZED AS
A SUBJECT IN QUEST OF A PREDICATE, AS A
BEING IN QUEST OF A MEANING OF LIFE, OF
ALL LIFE, NOT ONLY OF PARTICULAR AC-
TIONS OR SINGLE EPISODES WHICH HAPPEN
NOW AND THEN." (HESCHEL, JQ65, p. 54)

A proposition is a sentence. It is made up of nouns and a
predicate. Nouns are derived from signs; nouns can be con-
ceived as signs used in sentences. Verbs are not so easy to
characterize. Most verbs are also derived from signs; verbs
indicate actions instead of things. Adjectives and adverbs
also display this property of signification. Thus cow, green,
grass, run, chew, stand, trough, drink, water, are all signs
depicting events and occurrences. Only when used in sen-
tences do these signs become nouns, verbs, adjectives and
adverbs. What then makes a sentence?

Sentences are codes constructed by the mechanism of
predication. My hypothesis is that predication is a symbolic
process, i.e. it places linguistic signs into a context dependent
frame. Predication depends on the verb "is" in its various
grammatical constructions and according to my hypothesis
all basic sentences are explicitly or implicitly of the form
"X is Y."

As a corollary, predication is conceived to be a statement
of belief. (See Ayer, 1946, pp. 7-15 and 91-93, for similar
views.) The maker of a proposition is communicating his
belief with regard to a relationship among signs. Thus nega-
tion, qualification and the like are part of predication. The
sentence "the boy runs" is therefore a shorthand statement
of the sentence "the boy is running" and indicates certainty

OH the part of the speaker. "I believe the boy is running";
"I think the boy is running"; "the boy may or may not run*'
are all qualifiers on the certainty with whieh the proposition
is held. It is this process of making statements of certainty
of belief that is unique to man and provides the thrust to-
ward making experiences and encounters meaningful.

Propositions power meaning by introducing flexibility
into the relationship among signs. A new level of coding
emerges, the best formal example of which is the alphabet.
Each letter is a linguistic sign, a context-free indicator that
can be used as such — for instance, in organizing a diction-
ary. The symbolic use of the alphabet, on the other hand,
provides an infinite richness of meaning through combina-
tions of the self-same letters where context dependent rela-
tionships now become paramount. Thus "tap" and "pat"
have different meanings.

Man not only uses linguistic signs symbolically, he uses
linguistic symbols significantly. This he does when he rea-
sons. He takes a context dependent linguistic symbol and for
the duration of a particular purpose assigns to it a context-
free meaning. This is accomplished by making explicit a set
of rules governing the relationship among linguistic symbols
"for the duration." The set of rules is. of course, a set of
propositions. Algebra is probably the most familiar formal
example of reasoning.

The point at issue is that though animals make signs and
symbols, only man appears to use linguistic signs symboli-
cally in making propositions and linguistic symbols signi-
ficantly in reasoning. What then is different about man's
brain that makes possible a reciprocal interaction between
sign and symbol?

The common answer to this question is that man's brain
is characterized by its massive cortico-cortical connectivity

(Geschwind, 1965). This connectivity is conceived to be
quantitatively, not qualitatively, different from that of non-
human brains. But as we have already seen, the postulated
transcortical relay mechanism of sign and symbol construc-
tion does not come off well when examined in the light of
experimental evidence obtained with nonhuman primates.
Instead, signs and symbols are found to be made by virtue
of a mechanism that involves cortico-sw/?cortical connec-
tions that relay in structures hitherto conceived to be motor
in function. Thus if man's special capability is due to his
brain's cortico-cortical connectivity, this difference is quali-
tative not just quantitative.

The issue is an important one. If, in fact, the cortico-
cortical connectivity of man's brain proves to be the source
of his power of propositional language and reasoning, we
have an answer to the question of what makes man human.
A great deal is being made today of this cortico-cortical con-
nectivity in terms of the "disconnection" syndromes that
result in a variety of aphasias and agnosias. But data from
the clinic are not always easy to evaluate and misinterpreta-
tion due to unqualified preconceptions can readily occur.

I have some misgivings about the validity of the common
view that cortico-cortical connections are responsible for
man's human capabilities. I cannot now fully spell out these
misgivings because they are intuitive and constitute the ques-
tions directing my research plans for the immediate future.
But a few points can be made. Obviously the roots of the
misgivings lie in my experience with nonhuman brains.
Initially the cortico-cortical hypothesis seemed self-evident.
Only when experimental result after experimental result dis-
confirmed the hypothesis was I driven to search elsewhere
to make sense of the data. However, this is not all. The
cortico-cortical connection hypothesis implies that informa-

tion is transmitted ru the connections. The largest bundle of
connecting fibers, and one that has grown considerably in
size w hen man is compared to monkey, is the corpus callosum
which connects the two hemispheres. Yet this increase in the
connectivity between hemispheres in man has led to hemi-
spheric specialization, each hemisphere serving widely differ-
ent functions. The connections seem to make it possible for
the hemispheres to go their separate ways to a large extent
rather than to duplicate each other as they do in nonhuman
mammals (Pribram. 1962: Young, 1962).

Objections to this view of the functions of the corpus
callosum immediately come to mind as a result of Sperry's
( 1964 ) fascinating split-brain patients. Sperry demonstrates
that each hemisphere can be shown to control awareness in-
dependent of the other hemisphere once the callosum is cut.
He infers from this that separate consciousnesses, separate
minds, exist in one head in these patients. The assumption
underlying this inference is that ordinarily consciousness is
of a piece and that we are always single-minded. I challenge
this assumption. Single-mindedness is an achievement that
often demands considerable effort whether one is studying,
listening during a conversation, or driving an automobile.
Sperry's patients are not unique in being of two minds on
occasion.

Other evidence that gives rise to my misgivings with the
connectionist hypothesis comes from unilateral brain abla-
tions that produce symptoms which are alleviated by further
brain ablation. Thus unilateral ablations of the frontal eye-
fields in monkey and man result in a temporary disregard
of stimuli in the contralateral visual field ( Kennard. 1939:
Pribram. 1955 ). Such disregard does not occur if the lesion
is bilateralized. Also, unilateral occipital lobectomy in the
cat results in a homonymous hemianopia which is relieved

when the ipsilateral optic colliculus is removed (Sprague,
1966).

These are but straws in the wind but they prevent me
from obtaining too easy and early a closure on the problem
of what makes man human. In order that the issue can be
faced squarely, however, I must offer an alternative to the
cortico-cortical connection hypothesis. My alternative is
that man makes meaning through signs, symbols, proposi-
tions and reasoning by way of corticofugal-subcortical con-
nections that importantly involve the motor mechanisms of
the brain. I propose that man's thrust toward meaning de-
rives from the fact that his brain's motor mechanisms are
better developed than those of animals. These motor mech-
anisms are not to be conceived, as we have seen, merely as
movers of muscles. The brain's motor mechanisms are de-
vices that set the sensitivity of receptors and afferent chan-
nels, not just of muscle receptors but those of all receptors
(including eye and ear) as well. Changes in setpoint regu-
late awareness and behavior. The changes and their results
can relatively simply be encoded in brain tissue and thus
serve as guides subsequently.

Conclusion

THINKING 1^ 1 11 l\t. I\l> VO THOUGH! /^
BRED IS IN ISOLATED till l\ I III BRAIN.

(HESi in i /„„-,. p. si,

The implications for education of this propensity of the
brain for encoding and recoding its sensitivities are obvious.
In order to make information meaningful we must allow
pupils to encode in terms of their own sensitivities which
are not necessarily ours. They must be given the opportunity
to repeat the information given in such a way that it be-
:omes encoded in a context which makes meaning for them.
They must be encouraged to remake what we give them in
their own image.

This is not as difficult as it sounds. As already noted, even
young children who are deaf use signs differently from the
way Washoe the chimpanzee uses signs. Human children
spontaneously make propositions, their language is produc-
tive (Jakobsen. 1966). All neural tissue is spontaneously
active, nerve cells beat out electrical signals on their own
throughout life, much as does the tissue of the heart. In man
this spontaneity becomes organized early on so that he pro-
duces propositions, makes sentences. And then he begins to
play with these sentences, recoding them into different forms
and reasoning with them. Each new batch of teenagers
attests to the human proclivity for productively recoding
what is given. Why not utilize this marvelous capacity to
advantage in our educational effort?

To summarize briefly: man's brain is different in that it
makes imperative the productive use of linguistic signs
symbolically and linguistic symbols significantly. The flexi-
bility derived from this difference is immense. Given the

power of this flexibility man codes and recodes for fun and
profit. Every artistic endeavor, every working accomplish-
ment depends for its effectiveness not only on the informa-
tion conveyed by the theme but on the variations on that
theme. Human encounter is sustained not just by an ex-
change of information but by an infinite variety in familiar
communication. Animals use signs and symbols only in
special circumstances; man productively propositions all his
encounters and he reasons about all his experiences. Thus
man and only man shows this thrust to make meaningful
his experiences and encounters: he intends, he holds on to
his images.

But this is not all. By means of the motor mechanisms of
his brain man hopefully and continuously sets and resets
his sensitivities so that his images can become actualized in
his environment both by virtue of his own behavior and that
of socially contiguous others. Man's culture expresses these
hopes, this active thrust toward meaning. For to be human
is to be incapable of stagnation; to be human is to produc-
tively reset, reorganize, recode. and thus to give additional
meaning to what is. In short, "to be human is to be a prob-
lem." (Heschel, 1965, p. 105).

References Cited

Ayer, A. J.

1^46. Language, truth and logic. New York. Dover Publications.

BAD I Y. P., AND G. VON BONIN

1951. The isocortex of man. Urbana. University of Illinois Press.
Bernstein, n.

I l >67. The co-ordination and regulation of movements. New York, Pergamon
Press.

Bucy, P. C.

1949. The preeentral motor cortex. Second edition. Urbana, University of
Illinois Press.

Chomsky, N.

1963. Formal properties o( grammars. /// luce. R. D., R. R. Bush, and E.
Galanter (eds. ). Handbook of mathematical psychology. New York,
John Wiley & Sons, vol. 2.

Chow. K. L.

1952. Fuither studies on selective ablation of associative cortex in relation
to visually mediated behavior. Jour. Comp. Physiol. Psychol., vol. 45,
pp. 109-118.

1954. Lack of behavioral effects following destruction of some thalamic

association nuclei in monkey. Arch. Neurol. Psychiat.. vol. 71, pp.

762-771.
1970. Integrative functions of the thalamocortical visual system of cat. In

Pribram, K. H., and D. Broadbent (eds.), Biology of memory. New

York. Academic Press.

Dewson. J. H. Ill, K. W. Noble, and K. H. Pribram

1966. Corticofugal influence at cochlear nucleus of the cat: some effects of
ablation of insular-temporal cortex. Brain Research, vol. 2, pp.
151-159.

Evarts, E. F.

1967. Representation of movements and muscles by pyramidal tract neurons
of the preeentral motor cortex. In Yahr, M. D., and D. P. Purpura
(eds. ), Neurophysiological basis of normal and abnormal motor activi-
ties. Hewlett, New York, Raven Press, pp. 215-254.

Foerster, O.

1936. Symptomatologie der Erkrankurgen des Grosshirns: Motorische Felder
and Bahnen. /// Bumke, O., and O. Foerster (eds.), Handbuch der
Neurologic Berlin, J. Springer, vol. 6. pp. 1-448.

FOX, S. S., J. C. LlEBESKINDE, J. H. O'BRIEN, AND R. D. G. DlNGLE

1967. Mechanisms for limbic modification of cerebellar and cortical afferent
information. In Adey, W. R., and T. Tokizane (eds.). Progress in brain
research. Amsterdam, Elsevier Publishing Co., vol. 27, pp. 254-280.

Fritsch, G., and E. Hitzig

1870. On the electrical excitability of the cerebrum. Reprint. //* Pribram,
K. H. (ed.), Brain and behaviour. Baltimore, Penguin Books, vol. 2,
pp. 353-364, 1969.

Gabor. D.

1949. Microscopy by reconstructed wave fronts. Proc. Roy. Soc, London,
A 197, pp. 454-487.

1951. Microscopy by reconstructed wave fronts. II. Ibid., B64, pp. 449-469.
Galambos, R., T. T. Norton, and C. P. Frommer

1967. Optic tract lesions sparing pattern vision in cats. Exp. Neurol., vol. 18,
pp. 8-25.

Gardner, R. A., and B. T. Gardner

1969. Teaching sign language to a chimpanzee. Science, vol. 165, pp. 664-672.
Geschwind, N.

1965. Disconnexion syndromes in animals and man: part I. Brain, vol. 88,
pp. 237-294.

Granit, R.

1955. Receptors and sensory perception. New Haven, Yale University Press.
Granit, R., and J. O. Kellfrth

1967. The effects of stretch receptors on motoneurons. //; Yahr, M. D., and
D. P. Purpura (eds. ), Neurophysiological basis of normal and ab-
normal motor activities. Hewlett, New York, Raven Press, pp. 3-28.

Heschel. A. J.

1965. Who is man? Stanford, Stanford University Press.
Hubel, D. H., and T. N. Wiesel

1965. Receptive fields and functional architecture in two nonstriate visual
areas (18 and 19) of the cat. Jour. Neurophysiol., vol. 28, pp. 229-288.

Jakobsen, R.

1966. Linguistic types of aphasia. In Carterette, E. C. (ed.), Brain function.
Berkeley, University of California Press, vol. 4, pp. 67-91.

Jerison, H. J.

1961. Quantitative analysis of evolution of the brain in mammals. Science,
vol. 133, #3457.

Kamiya, J.

1968. Conscious control of brain waves. Psychology Today, vol. 1, pp.
56-60.

Kennard, M. A.

1939. Alterations in response to visual stimuli following lesions of frontal
lobe in monkeys. Arch. Neurol. Psychiat., vol. 41, pp. 1153-1165.

Kraft, M. S., W. D. Obrist, and K. H. Pribram

1960. The effect of irritative lesions of the striate cortex on learning of visual
discriminations in monkeys. Jour. Comp. Physiol. Psychol., vol. 53,
pp. 17-22.

KUFFLER, S. W., AND C. C. HUNT

1952. The mammalian small -nerve fibers: a system for efferent nervous
regulation of muscle spindle discharge. Res. Publ. Assoc. Nerv. Ment.
Dis., vol. 30, pp. 24-47.

Lashley, K. S.

1921. Studies of cerebral function in learning. The motor areas. Brain,

vol. 44, p. 255.

1950. in search of the engram. /// Society tor experimental biology (Gr. Mr. i

physiological mechanisms in animal behavior. New York. Academic
Press, pp. 454 482.

Lashley, K. S.. \m> G. Cl \kk

1945. I he cytoarchitecture ol the cerebral cortex of AnU-s: a critical
examination ol architectonic studies. Jour. (onip. Neurol., vol. 85,
PP. 223-305.

Leith, E. N.. \M> J. Upatnicks

1965. Photography b\ laser. Sci. Amer.. vol. 212, pp. 24-35.
I I hi i, B.

1966. Brain Stimulation and conscious experience. In Eccles. J. C. (ed.).
Brain and conscious experience. New York. Springer- Verlat:. pp.
165-181.

Mm is. L. I., K. H. Pribram, and L. Kri <,i k

1953. Action potentials in "motor" cortex evoked by peripheral nerve
stimulation. Jour. Neurophysiol.. vol. 16. pp. 161-167.

McCleary, R. A.

1961. Response specificity in the behavioral effects of limbic system lesions
in the cat. Jour. Comp. Physiol. Psychol., vol. 54. pp. 605-613.

Merton. P. A.

1953. Speculations on the servo-control of movement. In Malcolm, J. L.,
a^d J. A. B. Gray (eds.). The spinal cord. London. J. and A.
Churchill, pp. 247-260.

Mettler. F. A.

1967. Cortical subcortical relations in abnormal motor function. In Yahr.
M. D.. and D. P. Purpura (eds.). Neurophysiological basis of normal
and abnormal motor activities. Hewlett. New York. Raven Press.
pp. 445-497.

Miller. G. A.. E. H. Galanter. and K. H. Pribram

1960. Plans and the structure of behavior. New York. Henry Holt and Co.
Mishkin. M.

1969. Presidential address. Division of Physiological and Comparative
Psychology. American Psychological Association. Washington. D. C.
Sept. \-A.

Molntcastle. V. B.

1962. Interhemispheric relations and cerebral dominance. Baltimore, The
Johns Hopkins Press.

Premack. D.

1970. Testing a functional analysis of language with a chimpanzee. Invited
Paper, Western Psychological Association. Los Angeles, Apr. 17.

Pribram, K. H.

1955. Lesions of "frontal eye fields" and delayed response in baboons.
Jour. Neurophysiol., vol. 18. pp. 106-112.

1958a. Neocortical function in behavior. In Harlow. H. F.. and C. N. Woolsey
(eds.). Biological and biochemical bases of behavior. Madison. Uni-
versity of Wisconsin Press, pp. 151-172.

1958b. Comparative neurology and the evolution of behavior. In Simpson,
G. G. (ed.), Evolution and behavior. New Haven, Yale University
Press, pp. 140-164.

1961. Implications for systematic studies of behavior. In Sheer, D. E. (ed.),
Electrical stimulation of the brain. Austin, University of Texas
Press, pp. 563-574.

1962. Discussion in Mountcastle, V. B. (ed.), Interhemispheric relations
and cerebral dominance. Baltimore, The Johns Hopkins Press, pp.
107-111.

1964. Neurological notes on the art of education. //; Hilgard, E. (ed.),

Theories of learning and instruction. Chicago, University of Chicago

Press, pp. 78-110.
1966. Some dimensions of remembering: steps toward a neuropsychological

model of memory. In Gaito, J. (ed. ), Macromolecules and behavior.

New York, Academic Press, pp. 165-187.
1967a. The new neurology and the biology of emotion: a structural approach.

Amer. Psychologist, vol. 22, pp. 830-838.
1967b. The limbic systems, efferent control of neural inhibition and behavior.

In Adey, W. R., and T. Tokizane (eds.), Progress in brain research.

Amsterdam, Elsevier Publishing Co., vol. 27, pp. 318-336.
1969a. Four R's of remembering. In Pribram, K. H. (ed.), On the biology

of learning. New York, Harcourt, Brace and World, pp. 193-225.
1969b. The amnestic syndromes: disturbances in coding? In Talland, G. A.,

and N. C. Waugh (eds.), Psychopathology of memory. New York,

Academic Press, pp. 127-157.

1970. Feelings as monitors. In Arnold, M. (ed.), Feelings and emotions.
New York, Academic Press.

1971. Languages of the brain. New York, Prentice-Hall (in press).

Pribram, K. H., R. J. Douglas, and B. J. Pribram

1969. The nature of non-limbic learning. Jour. Comp. Physiol. Psychol.,
vol. 69, pp. 765-772.

Pribram, K. H., and L. Kruger

1954. Functions of the "olfactory brain." Ann. N. Y. Acad. Sci., vol. 58,
pp. 109-138.

Pribram, K. H., L. Kruger, R. Robinson, and A. Berman

1955- The effects of precentral lesions on the behavior of monkeys. Yale
1956. Jour. Biol. & Med., pp. 428-443.

Pribram, K. H., H. Lim, R. Poppen, and M. H. Bagshaw

1966. Limbic lesions and the temporal structure of redundancy. Jour. Comp.
Physiol. Psychol., vol. 61, pp. 368-373.

Pribram, K. H., M. Mishkin, H: E. Rosvold, and S. J. Kaplan

1952. Effects on delayed-response performance of lesions of dorsolateral
and ventromedial frontal cortex of baboons. Jour. Comp. Physiol.
Psychol., vol. 45, pp. 565-575.

Pribram, K. H., D. N. Spinelli, and S. L. Reitz

1969. The effects of radical disconnexion of occipital and temporal cortex
on visual behaviour of monkeys. Brain, vol. 92, pp. 301-312.

PRIBRAM, k, H . \M> I \\ I IskKVS 1/

1957. A comparison of the effects ol medial and lateral cerebral resections
on conditioned avoidance behavior of monkevs. Jour. Comp. Physiol.
Psycbol., vol. 50, pp. 74-80.

Pribkwi. k H w \ Wilson, and J. Connors

1 1 h>2. I he effects v\ lesions of the medial forehram on alternation behavior
of rhesus monkeys. Exp. Neurol., vol. 6. pp. 36-47.

<)i ii l i w \! k

1^67. Word concepts: a theoiv simulation of some basic semantic capabili-
ties. Behav. Sci.. vol. 12, pp. 410-^30.

Rtn/. S. I . ind K. H. Pribram

1969. Some subcortical connections of the inferotemporal gvnu of monkey.
E\p. Neurol., vol. 25. pp. 632-645.

Shi RRINCTON, C.

1906. The integrative action of the nervous svstem. Reprint. New Ha\en,
Yale University Press. 1947.

Sl'IKKV R. W.

1964. Problems outstanding in the evolution of brain function. James Arthur
lecture on the evolution of the human brain. New York. The American
Museum of Natural History.

SiMM i 1 1. 1) . \ . \M) K. H. Pribram

1966. Changes in visual recovery functions produced by temporal lobe
stimulation in monkeys. EEG Clin. Neurophysiol.. vol. 20, pp. 44—49.

Spinllli, D. N.. K. H. Pribram, and M. Weingarten

1965. Centrifugal optic nerve responses evoked by auditory and somatic
stimulation. Exp. Neurol., vol. 12. pp. 303-319.

Sprague, J. M.

1966. Interaction of cortex and superior colliculus in mediation of visually
guided behavior in the cat. Science, vol. 153. pp. 1544-1547.

Stamm. J. S.. and K. H. Pribram

1960. Effects of epileptogenic lesions in frontal cortex on learning and
retention in monkeys. Jour. Neurophysiol.. vol. 23. pp. 552-563.

1961. Effects of epileptogenic lesions in inferotemporal cortex on learning
and retention in monkevs. Jour. Comp. Physiol. Psvchol.. vol. 54,
pp. 614-618.

Stamm. J. S.. and A. Warren

1961. Learning and retention by monkeys with epileptogenic implants in
posterior parietal cortex. Epilepsia, vol. q. pp. 229-242.

van Heerden. P. J.

1968. The foundation of empirical knowledge. The Netherlands, Uitgeverij
W istik-Wassenaar.

Walshe, F. M. R.

1948. Critical studies in neurology. Baltimore. The Williams and Wilkins Co.

Woolsey. C. N.. H. T. Chang, and P. Bard

1947. Distribution of cortical potentials evoked by electrical stimulation of
dorsal roots in Macaca mulatto. Fed. Proc. vol. 6. p. 230.

Young. J. Z.

1962. Why do we have two brains? //; Mountcastle, V. B. (ed.). Inter-
hemispheric relations and cerebral dominance. Baltimore, The Johns
Hopkins Press, pp. 7-24.

Many important aspects of the problem of the brain's coding processes are
dealt with here altogether too briefly. But the present paper will serve as a
prolegomenon to a more comprehensive study which will appear under the title
Languages of the Brain: Experimental Paradoxes and Principles in Neuropsy-
chology, to be published by Prentice-Hall in 1971.

W-J - * V ■- • ^

FORTY-THIRD

JAMES ARTHUR LECTURE ON

THE EVOLUTION OF THE HUMAN BRAIN

19 73

THE ROLE OF HUMAN SOCIAL BEHAVIOR
IN THE EVOLUTION OF THE BRAIN

RALPH L.jHOLLOWAY

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1975

FORTY-THIRD

JAMES ARTHUR LECTURE ON

THE EVOLUTION OF THE HUMAN BRAIN

THE

FORTY-THIRD
JAMES ARTHUR LECTURE ON
EVOLUTION OF THE HUMAN BRAIN

1 '> 7 3

THE ROLE OF HUMAN SOCIAL BEHAVIOR
IN THE EVOLUTION OF THE BRAIN

RALPH L. HOLLOWAY

Professor of Anthropology, Department of Anthropology
Columbia University, New York

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1975

LIBRARY

OF THE

micoiriu iiiiccm

JAMES ARTHUR LECTURES ON
THE EVOLUTION OF THE HUMAN BRAIN

Frederick Tilney, The Brain in Relation to Behavior; March 15, 1932

C. Judson Merrick, Brains as Instruments of Biological Values; April 6, 1933

D. M. S. Watson, The Story of Fossil Brains from Fish to Man; April 24, 1934

C. U. Ariens Kappers, Structural Principles in the Nervous System; The Develop-
ment of the Forebrain in Animals and Prehistoric Human Races; April 25, 1935

Samuel T. Orton, The Language Area of the Human Brain and Some of its
Disorders; May 15, 1936

R. W. Gerard, Dynamic Neural Patterns; April 15, 1937

Franz Weidenreich, The Phylogenetic Development of the Hominid Brain and its
Connection with the Transformation of the Skull; May 5, 1938

G. Kingsley Noble, The Neural Basis of Social Behavior of Vertebrates; May 11,
1939

John F. Fulton, A Functional Approach to the Evolution of the Primate Brain; May
2, 1940

Frank A. Beach, Central Nervous Mechanisms Involved in the Reproductive
Behavior of Vertebrates; May 8, 1941

George Pinkley, A History of the Human Brain; May 14, 1942

James W. Papez, Ancient Landmarks of the Human Brain and Their Origin; May 27,
1943

James Howard McGregor, The Brain of Primates; May 11, 1944

K. S. Lashley, Neural Correlates of Intellect; April 30, 1945

Warren S. McCulloch, Finality and Form in Nervous Activity; May 2, 1946

S. R. Detwiler, Structure-Function Correlations in the Developing Nervous System
as Studied by Experimental Methods; May 8, 1947

Tilly Edinger, The Evolution of the Brain; May 20, 1948

Donald O. Hebb, Evolution of Thought and Emotion; April 20, 1949

Ward Campbell Halstead, Brain and Intelligence; April 26, 1950

Harry F. Harlow, The Brain and Learned Behavior; May 10, 1951

Clinton N. Woolsey, Sensory and Motor Systems of the Cerebral Cortex; May 7,
1952

Alfred S. Romer, Brain Evolution in the Light of Vertebrate History; May 21, 1953

Horace W. Magoun, Regulatory Functions of the Brain Stem; May 5, 1954

*Fred A. Mettler, Culture and the Structural Evolution of the Neural System; April
21, 1955

*Pinckney J. Harman, Pale one urologic, Neoneurologic, and Ontogenetic Aspects of
Brain Phytogeny; April 26, 1956

*Davenport Hooker, Evidence of Prenatal Function of the Central Nervous System in
Man; April 25, 1957

*David P. C. Lloyd, The Discrete and the Diffuse in Nervous Action; May 8, 1958

*Charles R. Noback, Tfie Heritage of the Human Brain; May 6, 1959

*Ernst Scharrer, Brain Function and the Evolution of Cerebral Vascularization; May
26, 1960

Paul I. Yakovlev, Brain, Body and Behavior. Stereodynamic Organization of the
Brain and of the Motility -Experience in Man Envisaged as a Biological Action
System; May 16, 1961

H. K. Hartline, Principles of Neural Interaction in the Retina; May 29, 1962

Harry Grundfest, Specialization and Evolution of Bioelectric Activity; May 28, 1963

*Roger W. Sperry, Problems Outstanding in the Evolution of Brain Function; June 3,
1964

*Jose M. R. Delgado, Evolution of Physical Control of the Brain; May 6, 1965

Seymour S. Kety, Adaptive Functions and the Biochemistry of the Brain; May 19,
1966

Dominick P. Purpura, Ontogenesis of Neuronal Organizations in the Mammalian
Brain; May 25, 1967

*Kenneth D. Roeder, Three Views of the Nervous System; April 2, 1968

f Phillip V. Tobias, Some Aspects of the Fossil Evidence on the Evolution of the
Hominid Brain; April 2, 1969

*Karl H. Pribram, What Makes Man Human; April 23, 1970

Walle J. H. Nauta, A New View of the Evolution of the Cerebral Cortex of
Mammals; May 5, 1971

David H. Hubel, Organization of the Monkey Visual Cortex; May 11, 1972

Janos Szentagothai, The World of Nerve Nets; January 16, 1973

* Ralph L. Holloway, The Role of Human Social Behavior in the Evolution of the
Brain; May 1, 1973

*Elliot S. Valenstein, Persistent Problems in the Physical Control of the Brain; May
16, 1974

*Published versions of these lectures can be obtained from The American Museum of
Natural History, Central Park West at 79th St., New York, N.Y. 10024.

fPublished version: The Brain in Hominid Evolution, New York: Columbia
University Press, 1971.

THE ROLE OF HUMAN SOCIAL BEHAVIOR
IN THE EVOLUTION OF THE BRAIN

INTRODUCTION

The presentation of this lecture has particular significance for
me because only slightly more than 11 or 12 years ago as a
graduate student of human evolution I discovered with great
excitement the existence of the James Arthur lectures; these
surely decided my fate, at least in part.

I wish to discuss some of the brain endocasts of our earliest
fossil hominid ancestors and to show that the human brain has
been around for quite a long time, perhaps three million
years — or longer. This is somewhat of an about-face for me, for
when I wrote my dissertation about 10 years ago, I regarded
endocasts as so much rock or plaster, with little, if any,
potential of offering evidence on the evolution of the hominid
brain. I believe I have mellowed. Today endocasts are the
subject of my major research effort.

My questions about the human brain are: What lines of
evidence can we use to learn about it; how did it evolve to its
present state; can we find something in its evolution relevant for
today's societal existence?

Before I discuss these questions in detail, I wish to briefly
consider my basic conclusions:

1 . The usual orthodox version of hominid evolution places the
evolution of the brain as a terminal phase, one that occurs
after all other parts of the body, such as the hands, the trunk,
the teeth, and the locomotory anatomy for bipedalism have
evolved. This view is very oversimplified, if not downright
incorrect, and approximates truth only if we are willing to
equate brain evolution with brain enlargement. Indeed, the
evidence shows that brain modification to a human pattern
occurred early in human evolution, at least three million
years ago.

2. Both brain endocasts and bodily skeletal parts suggest that
brain :body relationships remained fairly constant during
most of human evolution, indicating an important set of
selection pressures for body-size increase. This evidence also
suggests that brain encephalization, as measured by Stephan's
(1972) "progression indices" (related to a "basal insectivore"
line), was already within the human range in the early fossil
hominids. The mediating factor for increase must have been
an endocrine-target tissue adjustment resulting in selection
for increased delays of maturation, or prolonged growth and
dependency times, important factors in any consideration of
social behavior.

3. The humanly organized brain and resultant human cultural
behavior have been interacting in a positive feedback manner
during most of human evolution (Holloway, 1967). This
feedback interaction is probably over, and unless some new
radical genetic change occurs to interrupt man's present
growth pattern, or a new social order that practices some
form of genetic surgery comes into existence, I do not believe
the human brain will show any further significant evolution
in terms of size increase.

4. Brain endocasts have enormous value in the study of human
evolution that extends far beyond brain-behavior correla-
tions. They can give us information about variation, popula-
tion statistics, and brain :body ratios, and therefore have
importance in relating early hominid populations to ecologi-
cal parameters such as biomass and growth and development.

5. Finally, we must realize that human behavior is not a recent
achievement — our social behavior, our sociality has long evo-
lutionary roots that cannot be abridged simply by cultural fiat.

Abbreviations used in the text and figures are:

ER, East Lake Rudolf

HE, Indonesian Homo erectus

MLD, Makapansgat, S. Africa

OH, Olduvai Gorge

OMO, Omo Valley, Ethiopia

SK, Swartkrans

STS, Sterkfontein

LINES OF EVIDENCE
DIRECT

It lias long been appreciated that the only direct evidence for
the study of brain evolution comes from the endocasts of our
fossil ancestors (Edinger, 1929, 1949, 1964; Holloway, 1964,
1966a; Radinsky, 1967, 1970). Whether they are natural
endocasts of the South African australopithecines (e.g., Taung,
STS 60, Type 3, and SK 1585) or prepared in the laboratory
from latex, plaster of Paris, and plasticine, they give only the
most limited information about neural structure and no direct
information about behavior. An endocast is simply a mold of
the inside bony table of the cranium. Between the bone and the
underlying brain there are three meningeal tissues of varying
thickness, as well as a variably distributed amount of cerebro-
spinal fluid. The thick dura mater, the arachnoid space, the
investing thin layer of pia mater, and the cerebrospinal fluid all
"conspire" to eradicate the sulcal and gyral configurations
imprinted by the surface of the cerebral cortex into the bony
layer of the cranium. This "conspiracy" varies in different
orders of animals; it is most severe, unfortunately, in the living
and fossil species of apes and man. The reasons for this and the
reasons for variation with age are not totally understood, but
they are probably linked to differential growth rates of the
brain and the overlying cranial bones in different regions (e.g.,
Hirschler, 1942; Keith, 1931).

Endocasts can be obtained from fossil cranial fragments in
two ways. Natural endocasts occur when the skull is filled by
fine sediments drifting through the cranial foramina, particu-
larly the foramen magnum. The sediments may be compacted
and solidified by percolating mineral solutions, resulting, in
time, in a solid mass of sedimentary rock inside the skull. The
skull bones may eventually erode away leaving the endocast
intact. Usually the skull is preserved around the endocast, as is
sometimes the case with the South African australopithecines,
such as the Taung specimen, STS 60, Type 3, and the more

recent SK 1585 (figs. 1-8). In SK 1585 I deliberately removed
the already eroded bones to disclose the fine-grained natural
endocast (see Holloway, 1972a for details).

Endocasts may also be made by applying liquid rubber latex
to the inner cranial surface of a skull. This method has been
used for most of the endocasts, including all the rest of the
hominids from East Africa, Asia, and Europe. Successive layers
are built up until a reasonable thickness, perhaps an eighth of an
inch, is reached. The latex is cured by heat and then collapsed
from the skull, either before or after stabilizing the dimensions
with plaster. The external details of the cerebral cortex, as
transmitted through the dura mater, will be reproduced on the
surface of the latex. If the inner bony table is eroded before the
endocasts are made, the details will obviously be missing.

"V;

Ti — G — I

FIG. 1. Lateral view of Taung infant endocast and face positioned together.
Arrow points to lambdoid suture, which is probably the most anterior extent of
lunate sulcus. Scale equals 3 cm.

FIG. 2. Lateral view of Taung infant endocast. Arrow points to third inferior
frontal convolution. Small portion of frontal lobe remains embedded in facial
fragment. Scale equals 3 cm.

INDIRECT

Brains influence behavior, and occasionally the results of
behavior become, so to speak, fossilized. Fortunately, the
paleoanthropologist has lines of evidence for the evolution of
the brain other than brains or endocasts. There are two sources
of indirect evidence: (1) cultural products of brain and social
behavioral activity, e.g., stone tools, shelters, animal remains at
ancient butchering sites; and (2) skeletal components of the
masticatory and locomotor systems. No indirect evidence can
yet be used to demonstrate any specific changes in the brain
observable at the surface. It is, however, indicative of different
behavioral capabilities, which require, after all, neural com-

FIG. 3. Occipital view of Taung infant endocast. Lambdoid suture is distinct.
Notice gyral curvature (shown by dotted line and arrow) immediately superior and
anterior to lambdoid suture, indicating that more forward placement of lunate sulcus
would not be possible. Scale equals 3 cm.

plexes to effect them. In other words, it supports the idea of
brain reorganization.

The first line of indirect evidence applies, as far as we know,
only to hominids. There is no evidence from the fossil record of
the cultural behavioral effects in other lines of primates. The
second line of indirect evidence, that is, musculoskeletal, is far
more general and applies to all lines of animals, most particu-
larly to the mammals. But what we see in the hominid fossils is
rather specific, at least when compared with other fossil
primates, or extant ones, for that matter. The earliest hominids
show definite changes in masticatory apparatus — in the teeth,
jaws, and areas of muscle attachment for the temporalis and

FIG. 4. Lateral view of plaster replica of SK 1585, endocast from Swartkrans,
South Africa. A small portion of frontal lobe is missing. Lambdoid suture obscures
posterior limit of lunate sulcus. (See Holloway, 1972a.) Scale equals 3 cm.

masseter in particular. We find changes in the molars as far back
as 10 to 14 million years ago in Ramapithecus (Pilbeam, 1969;
Simons, 1961, 1964, 1969). Among the early hominids of East
and South Africa there are changes in nuchal musculature
related in part to advanced degrees of bipedal locomotion,
which itself is corroborated by the remains of the locomotor
skeleton (pelvis, lower vertebral column, limb bones such as the
femur, tibia and fibula, and various bones of the foot). Even the
hand bones, at least of the East African hominids, show changes
in musculoskeletal structure suggestive of manipulative abilities
greater than those of any fossil or living ape or monkey.

Why belabor these points? Because they show, whether or
not the precentral gyrus appears on the surface of the endocast,

FIG. 5. Occipital view of plaster replica of SK 1585 endocast. Scale equals 3 cm.

that natural selection has long been operating on behavior,
favoring neural organizations capable of servicing the new
musculoskeletal complexes.

This line of indirect evidence for brain reorganization need
not be related only to motor or sensorimotor behavior, such as
the various muscle contractions involved in bipedalism, but it
must be taken to involve the whole adaptive complex (hunting,
scavenging, carrying objects, and so on) in which these motor
patterns are embedded and to include aspects of psychological
restructuring as well. It is true that there is yet no way of
comparing a gorilla endocast to that of an australopithecine or a

FIG. 6. Basal view of plaster replica of SK 1585 endocast. Scale equals 3 cm.

Homo sapiens to show correlated changes between brain surface
features and motor behavior. Endocasts may or may not reflect
important adaptive changes in behavior and structure, but by
themselves they cannot indicate whether the brain evolved
before or after the sensorimotor changes.

THE EVIDENCE

Evidence from which I conclude that the brain has always
been an important component of human evolution is as follows:

1. Gross Morphology: Hominid endocasts show a human
shape that is not found among a sample of 50 chimpanzee and
gorilla endocasts. Although there can be considerable variation
in endocasts of living pongids (figs. 9-16), none shows the
combination of features seen on hominid endocasts. The
differences are as follows:
a. The height of the brain above the cerebellar lobes is almost

always greater in hominid brains. Occasionally the brain of

FIG. 7. Dorsal view of Type 3 endocast, a gracile australopithecine from
Sterkfontein. Note double-valleyed fracture in parietal lobe, squared-off shape of
frontal lobe, and suggestion of heavy gyral and sulcal relief. (See Schepers, 1946.)
Scale equals 3 cm.

the pygmy chimpanzee, Pan paniscus, shows less flattening in
height than that of either the gorilla or the chimpanzee (Pan
troglodytes sp.) but it is not so high as that of the early
australopithecines (table 1).
b. The anterior tips, or poles, of the temporal lobes are

FIG. 8. Dorsal view of endocast of STS 60 from Sterkfontein. Scale equals 3 cm.

distinctly more rounded and larger in hominids than in
pongids (part of this is, of course, due to the different shape
of the greater wing of the sphenoid and the dural sheath
surrounding the tip of the lobe).

c. The orbital surface of the frontal lobe is generally angled
upward, with a more pointed and pronounced beak in pongid
than in hominid brain casts.

d. In pongid endocasts the position of the famous "lunate" or
"simian" sulcus, which divides the primary visual cortex from
the so-called parietal "association" cortex, is usually in a
fairly anterior position (although less so than in cerco-
pithecoids). Although only a few hominid endocasts [particu-
larly the original Taung (1924) endocast] show the sulcus

clearly, it is definitely in a posterior, human-like position
(figs. 3, 5). It is probably this feature, more than any other,
that so firmly suggests cortical reorganization to a human
pattern. This observation was first noted by Dart (1925),
later by Schepers (1946), and was more or less verified by Sir
Wilfred LeGros Clark (1947); a close examination shows no
alternative position.

e. The inferior border of the temporal lobe also shows
enlargement, reflected in a smaller, or more acute, angle of
the petrosal cleft.

Taken together, these features form a Gestalt that is very
difficult to demonstrate by linear measurements, as many
physical anthropologists would wish. It is these 'Gestalten'
that enable one to distinguish between pongid endocasts,
such as between those of chimpanzees and gorillas, even
though most measurements and indices tend to overlap.

f. Finally, it is possible that there is more sulcal and gyral
development in hominid cortices, particularly on the frontal
lobe, than in pongid cortices; however, this is not easily
measured on endocasts and is at best an impressionistic
judgment.

2. Gross Size: This parameter (or to follow Jerison, 1973,
"statistic") is perhaps the crudest of all. The small absolute sizes
of the australopithecine endocasts tended to deny them
hominid status long after their discovery. Elsewhere I (Hollo-
way, 1964, 1966a, 1968, 1970, 1972b) have detailed my
observations on the signficance of this measurement of the
brain. Some chimpanzees and most gorillas have larger brains
than the early hominids (see, for example, Tobias's 1971
compilations). The range of variation in normal present-day
Homo sapiens is from about 1000 to 2200 cc, or about as
much as the total evolutionary gain from Australopithecus
africanus, at ca. 450 cc, to the average value of modern Homo
sapiens of about 1400 cc. Yet there has never been any
demonstration, among living populations, of a relationship
between brain size (measured either by weight or volume) and
behavior. Although some human microcephalics have brain

FIG. 9. Lateral views of rubber latex endocasts of (top) Pan paniscus, pygmy
chimpanzee, (middle) Pan troglodytes, and (bottom) Gorilla gorilla. (Rubber latex
endocasts made by author from specimens belonging to the American Museum of
Natural History.) (See figs. 10 and 11 for occipital and dorsal views of same
specimens.) Scale equals 3 cm.

■a

-2

^-H

E3 E

M *

D 0>

S 2

— i-i
CO ^

p— I CT\

[J- .

—

■a

3
—

7^;

4>
U

—

2 o

•§3

4) ^

-a -*

. o

FIG. 13. Dorsal view of endocast of different, or another Pan troglodytes
showing excellent gyral and sulcal markings. Arrows indicate anterior limit of lunate
sulcus. (Rubber latex endocast made by author from a specimen at the American
Museum of Natural History.) Scale equals 3 cm.

volumes that gorillas, and perhaps a few large chimpanzees,
might disdain, they do not exhibit simian behavior, but rather
show the species-specific ability for symbolic language, albeit
disadvantaged.

The usefulness of this crude measure of the brain lies in its
statistical utilization as a parameter from which other neural
measures, such as neuron size, glial/neuron ratio, neural density,
and dendritic branching may be calculated. All of these
variables are closely tied in with behavioral variation, although
it remains for future scientists to demonstrate this unambigu-
ously (see Holloway, 1964, 1966a, 1966b, 1968; Jerison,
1973).

FIG. 14. Endocast of modern Homo sapiens, lateral view. Note great height of
cortex above cerebellum, expansion of temporal lobe in anterior and posterolateral
margins, and slight slope of orbital surface on frontal lobe. (Rubber latex endocast
made by author from specimen belonging to Columbia University.) Scale equals 3 cm.

Gross brain size is also related to body mass and time, and
thus it can be used in combination with these other variables to
give us clues about changes in growth rates during evolution in
particular phyletic lines. The study of brain -.body allometric
relationships in different animal lines has had a long history and
is a subject that is receiving considerable attention by modern
scientists (see, for example, Jerison, 1973). So far, however,
most of these studies have been concerned with comparisons
between high-level taxa, such as between carnivores and
herbivores, reptiles and birds, pongids and modern man. But if
brain and body size can be measured with reasonable accuracy
within a phyletic line, such as the Hominidae, the changes in
allometric relationships with time can provide extremely im-

FIG. 15. Same specimen as figure 14, occipital view. Scale equals 3 cm.

portant clues to selection pressures operating on variables such
as growth rates of different parts of the body, encephalization,
postnatal growth, and so on, which obviously have important
biological relationships with social behavior and adaptation. In
other words, another significant use of gross brain size, beyond
that of simply indicating overall size increase, is as a key to
other relationships that may have been more concerned with
selection pressures.

Unfortunately our samples for various hominid lineages are
terribly small, and many specimens (e.g., the South African
hominids) are not firmly dated; it is thus impossible to plot
brain size against time in any accurate manner. If we could, the
rates might give us some interesting clues to past selection

FIG. 16. Same specimen as figure 14, basal view. Scale equals 3 cm.

pressures and dynamics (see Holloway, 1972b). Table 2 gives a
number of newly determined endocranial capacities for various
hominids. The methods used to arrive at these figures are given
in the footnotes to this table.

3. Relative Brain Size and Encephalization: There appears
to be a lawful relationship between brain and body size in all
vertebrate taxa (see Jerison, 1973, for a thorough review of this
relationship). In general, following the principle of allometry,
larger-bodied animals tend to have a proportionally smaller brain
weight. It is possible to plot the size of the brain against the
weight of the body on double-logarithmic graph paper and to
discern some reasonably straight-line relationships. Regression
lines are of the general form E = kP^, where E = brain weight,
P = body weight, y = an exponent probably reflecting the

relationship between volume and surface area, and k = a
constant, often taken to reflect "encephalization," or the

relationship between brain :body weight ratios in different
animals. Plotting different orders of vertebrates on the same
graph tends to give an exponent of 0.66; for elosely related
species the exponent usually falls between 0.20 and 0.30.
Within the speeies, however, there seldom appears a relation-
ship, but this is probably debatable. 1 The human brain is neither
the smallest nor the largest in terms of relative size. Table 3
gives a few examples of animals with large and small relative
brain weights. This table does not show, however, the range of
variation within each category for brain: body weight ratios, for
which few published data exist.

Using a large number of "basal" insectivores (representing the
sort of primitive stock out of which the primates may have
evolved), Stephan (1972) was able to construct a "basal"
insectivore line, defined as log 10 h = 1.632 + 0.63 log 10 k. By
substituting a primate's body weight in the equation (k), it is
possible to solve for "h," which gives the expected brain weight
of a "basal" insectivore with such a body weight. If this weight
is then divided into the actual brain weight of the particular
primate, an "index of progression," or measure of encephaliza-
tion, results (table 4 shows a number of "progression indices"
for different primates, including some fossil hominids). This last
step requires making a hazardous assumption about the body
weight of the fossil hominid. Nevertheless, allowing for maximal
and minimal body weight, the South African gracile australopi-
thecines fit either within the range for modern man or just
below it, but always above the pongid range. This is indirect
evidence for reorganization of these early hominid brains to a

'Very little secure data exist for large samples of healthy individuals, which
requires study by more sophisticated statistical methods, such as partial correlations.
To date no such study has been published, not even in the excellent article by
Pakkenberg and Voigt (1964) on the Danes. I give this warning because a preliminary
analysis of a partial correlational study between the variables of age, weight, body
height, and brain weight suggests more of a relationship between brain and body
weight than is usually recognized. I hope to publish these results in the near future,
thanks to the courtesy of Dr. Pakkenberg, who has given me the original data.

human pattern, but it does not tell us whether there is a general
allometric increase in overall size or whether there has been
differential development of particular elements of the brain.

Still, these data are more relevant for understanding evolu-
tionary change than are mere comparisons of gross brain size. It
is a great pity that we do not as yet have a way to determine
accurately the body weights of our hominid ancestors. If we did,
we could plot these for particular lineages and, possibly, relate
the resulting exponents to evolutionary selection pressures.

Figure 17 and table 5 show a range of possible brain:body
weight relationships, based on current estimates of hominid
body weights (Tobias, 1967; Lovejoy and Heiple, 1970) that
might have characterized stages of hominid evolution. Inter-
pretation of selection pressures for increasing brain size varies,
depending on whether the exponents linking the fossil hominid
lineages are > 1.0, 1.0, 0.66, or less. The exponent 0.66
characterizes most nonhominid mammals (Jerison, 1973),

TABLE 1

Some Crude Indices for Hominid Endocasts a

Volume in

Dare

H 3

Specimen

Milliliters

Larc

Taung

404

1.13

1.48

1.41

STS60

428

1.00

1.35

1.40

1.29

STS5

485

1.08

1.39

1.42

1.27

OH 5

5 30

1.47

1.37

1.45

1.20

SK 1585

530

1.73

1.42

1.43

1.37

ER732

506

1.06

1.42

1.48

1.13

OH 24

590

1.01

1.29

1.40

1.32

OH 13

650

1.17

1.49

1.48

1.16

OH 9

1067

1.05

1.31

1.55

1.18

OH 12

727

1.11

1.41

1.60

0.97

HE I 6

943

1.10

1.33

1.59

1.02

HE II 6

815

1.06

1.35

1.53

1.08

HE IV 6

900

1.00

1.31

1.64

0.94

HE VI 6

855

1.05

1.33

1.68

0.97

HEVII &

1059

1.07

1.4!

1.65

0.92

HE VIII 6

1004

0.98

1.25

1.61

1.00

ER 1470 c

770

1.04

1.37

1.36

1.30

Omo 338s

427

1.02

1.37

1.54

1.03

TABLE ! - (Continued)

Volume m

I) arc

1) arc

H 3

Specimen

Milliliters

L arc

Pan paniscus

<n = 8)

average

325

0.99

1.33

1.46

1.04

range

284-363

0.97-1.01

1.28-1.37

1.36-1.54

0.86-1.21

Pan troglodytes

(n = 29)

average

394

0.96

1.28

1.47

1.09

range

334474

0.88-1.01

1.20-1.34

1.39-1.59

0.95-1.23

Gorilla gorilla

(n= 36)

average

498

0.98

1.26

1.53

1.04

range

383-625

0.94-1.04

1.19-1.33

1.39-1.67

0.85-1.24

Homo sapiens

(n = 4)

average

1442

1.10

1.43

1.40

1.25

range

1324-1586

1.04-1.14

1.39-1.46

1.35-1.46

1.11-1.42

Symbols: D arc = dorsal measurement between frontal and occipital poles; L arc =
lateral measurement between frontal and occipital poles; L = chord length between
frontal and occipital poles; H = chord length from vertex to lowest plane of temporal
lobe; V = volume.

a These figures clearly show that most hominid fossils (Homo erectus excepted) have
a greater degree of cortical height relative to both length and volume than do the
African pongids tested.

ft The well-known platycephalyof the Indonesian H. erectus is clearly shown by the
L/H value, the low D arc/L arc, D aic/L and H 3 /V ratios.

c This specimen does not show a typical H. erectus pattern.

"basal insectivores," and most lower primates (Stephan, 1972).
This exponent suggests an allometric increase, where brain
weight increases at a smaller rate than body weight. An
exponent of approximately 1.0 indicates a constant brain :body
weight ratio, suggesting selection pressure for brain weight to
match body weight. An exponent greater than 1 .0 suggests selec-
tion pressures for brain weight greater than that for body weight. 1

'Of course, an exponent of 1.0 in hominids does mean an increase in brain size
when compared with either a "basal" insectivore or vertebrate line where the
exponent is about 0.66.

TABLE 2
Endocranial Volumes of Reconstructed Hominid Specimens

Endocranial

Volume in

Specimen

Taxon

Region

Milliliters

Method

Evaluation

Taung

A. africanus

South Africa

440 c

STS 60

A. africanus

South Africa

428

STS71

A . africanus

South Africa

428

2-3

STS 19/58

A . africanus

South Africa

436

STS 5

A. africanus

South Africa

485

MLD 37/38

A. africanus

South Africa

435

MLD 1

South Africa

500±20

SK 1585

A . robustus

South Africa

530

OH 5

A . robustus

East Africa

530

OH 7

H. habilis

East Africa

687

OH 13

H. habilis

East Africa

650

OH 24

H. habilis

East Africa

590 rf

2-3

OH 9

H. erectus

East Africa

1067

OH 12

H. erectus (?)

East Africa

727

2-3

ER406

A . robustus

East Africa

510±10

ER732

A . robustus

East Africa

500

ER 1470

H. sp.?

East Africa

770 e

HE1

H. erectus

Indonesia

953^

HE 2

H. erectus

Indonesia

815-f

HE 4

H. erectus

Indonesia

90(/

2-3

HE 6 (1963)

H. erectus

Indonesia

855^

HE 7 (1965)

H. erectus

Indonesia

1059-f

1-2

HE 8 (1969)

H. erectus

Indonesia

1004/

a A, direct water displacement of either a full or hemiendocast with minimal
distortion and plasticine reconstruction; B, partial endocast determination, as
described by Tobias (1967, 1971); C, extensive plasticine reconstruction, amounting
to half the total endocast; D, determination based on the formula V = f Vz (LWB +
LWH), described by MacKinnon et al. (1956), where L = maximum length, W =
width, B = length, bregma to posterior limit of cerebellum, H = vertex to deepest part
of temporal lobe and f appears to be a taxon specific coefficient.

ft An evaluation of 1 indicates the highest reliability, 3, the lowest.

Postulated for adult-the value of the actual specimen is 404 ml.
Possible overestimate.

Provisional estimate.

■^These values are as yet unpublished and should be regarded as provisional.

At the present stage of our knowledge, it is premature to go
beyond this kind of simple exercise. Our samples are extremely

small, we have no good empirical evidence for any early
hominid body weight and the values in figure 17 connect
lineages that are geographically separated (i.e., the South
African gracile Australopithecus with the East African Ilahilis
with the East Asian Homo erectus with modern Homo sapiens).
Nevertheless, these relationships between brain and body weight
hold great promise for better understanding the dynamics of
hominid evolution. Indeed, as is clear from figure 17, one can
draw the lines in different ways, with constant slopes (i.e., 1.0)
or with different slopes at different times. (See also Holloway,
1974a.) The implications are extremely important, even though
the basic data are admittedly weak, for the lines in figure 16
demonstrate that a number of alternative hypotheses about
hominid brain evolution can exist, and that any particular
hypothesis is based on assumptions of body weight that cannot
be empirically pinpointed. In any event they do show a human,
rather than a pongid, pattern in terms of relative brain size and
changes through time, which strongly suggests that hominid
brain size increase and attending selection pressures were
probably unique.

TABLE 3

Some Average Brain:Body Ratios for Various Animals

Brain :body
weight ratio

Homo sapiens b

Gorilla

Chimpanzee

Macaque, Rhesus

Marmoset

Squirrel monkey

Elephant

Whale

Porpoise

200

185

170

600

10,000

"From Cobb, 1965.

"Good tabulated data on ranges for healthy human adults is lacking. The excep-
tion is one study on Danes by Pakkenberg and Voigt, 1964, p. 297, in which normal
brain:body weight ratios are shown to vary from approximately 1:28 to 1:80.

TABLE 4
Some Possible Brain Size: Body Weight Ratio and "Progression Indices"

Brain :

"Progression

Average Brain

Assumed Body

Body

Index"

Specimen

Size (ml.)

Weight (Pounds)

Ratio

PRG/BG

Gracile australopithecine

442

1:41

21.4

18.7

16.9

Robust australopithecine

530

22.3

19.9

16.9

110

12.8

Homo sapiens

1361

150

28.8

Homo erect us

930

26.6

125

22.0

"Based on Stephan's, 1972, formula using a basal insectivore line (see text).

Note: Maximum PRG/BG for gorilla is about 7.0, and for the chimpanzee, about
12.0. See Stephan, 1972, for ranges.

Maximum body weight of average gracile australopithecine (442 ml.) with
PRG/BG of 12, is 100 pounds. That is, if we allow the "progression index" of
Australopithecus to be the maximum chimpanzee value, the body weight is calculated
to be 100 pounds, which is clearly too heavy, based on the postcranial materials we
have thus far discovered for the gracile form of Australopithecus.

4. Lateralization and Cerebral Hemispheric Dominance:
Comparative neuroanatomy has not been able to demon-
strate any definite difference between the human brain and
the ape brain except on the basis of size. Absolute and relative
brain sizes, plus quantitative differences in amount of cerebral
cortex in certain lobes, such as the parietal and temporal, are all
that have been defined. These are matters of continuity, as far
as can be established at present. The cortico-cortical fasciculus
occipito-frontalis, a long associational tract known to exist in
the human brain, has not been distinguished in the chimpanzee
or cercopithecoid brain (Bailey et al., 1943). This does not
mean that an associational system does not exist between the
posterior and frontal segments of the chimpanzee cortex, but
only that it is probably not so developed as it is in Homo
sapiens. Many more pongid specimens should be dissected
before its presence or absence can be proved.

TABLE 5

Brain:Body Weight Double-Log Relationships Based on the General F-ormula h = bk x ,

with Possible Slope Differences Depending on Brain and Body Weights Used 17

Brain Volume

Body Weight

Specimen

(ml.)

(pounds)

Slope

A ustralopithecus

450

1.0

H. erectus

930

103

123

1.0

H. sapiens

1361

123

150

180

Australopithecus

450

0.6

H. erectus

930

115

143

180

0.6

H. sapiens

1361

200

250

Australopithecus

450

85-86

1.92

H. erectus

930

123

1.92

H. sapiens

1361

150

A . africanus

450

1.0

H. habilis

775

—

0.6

H. erectus

930

114

—

1.75

H. sapiens

1361

140

a Brain weights are held constant, the slopes varied and the resulting body weights
determined by projection to the abscissa! axis, which is the body weight.

The most singular difference known to exist at present is that
the human brain is characterized by cerebral hemispheric
dominance and a high degree of laterality. In general the left
hemisphere seems dominant, in terms of language phenomenon,
in the inferior parietal lobe (Wernicke's area), in the gyri and
sulci of Heschl and in the third inferior frontal convolution,
often known as Broca's area. The right hemisphere, particularly
the parietal lobe, seems "dominant" for spatiotemporal and

(J)

CD
CD

»-

h-
_l

_l
<

CO
1—

1-
_l

h-
_l

Q-

•

>
a

o
m

CD
O

(lH9l3MNIVd9) OL D01

v. C

50 —

C u-

o x a

"O II II

. X x — • _

C 3 B _ -a bo

~ u. r: O n —

.2 .- g 1 1

.2 J

C <u c

a c -
o o *-

as a

' X 3" § I I

WOO) *"

£<* * 2-= g.

n O >, ft +- g

(U gd " -— ' P 8

§ a* . _ — o

& <U « qvO*

c £ >/^ f> ***

•£ ^ M =/j T3 m

* 2 o « S «

2 M '■! c "H

"S -■> X 5> "^ n
.2 M ^ 3 M

»* "9 :s ^ u

I 1 c .SO -

H o £ —

^ j c

o -d ii o

c3 x x

o » S "S.
S x 2 o

XI so — — -

y * i: o o

60 *

2 ^°

« o

*§tl

-r >>

« c

. X,- = .SO

B "d - _ 3 —

— o 60 «j £ w

S x> 5 "3 5 <

5 — ^

a o +

• V5

P~ C X)

- .2 =

■->

. > 2

S J JS

f § i

o ~ c

u 53 *-"

D. e9

a 5-a

5 S |

60 "^

■£ >. £

60 q J> XJ M _
ill n ** "^ rti "2

i> X

>> - o 53

o x w
x H 8

2 Q.<

< £ S

3 S .5 8

<u iz
x n

— T3

C <U „

3 s S

o 2 X

Q. £ J,

6 » jj

a •£ c

so rt

C T3

O 3

5 £

60 -O

Q "S S »

_ — X —

2 & o" i» l ^ i

3 11 o a 3 g«

rj X < 3 H 00

3 E o

T3 '■" —
C 2 ji

« x c

- .»-a

ugh

X S -°

I o 'i

*» X 4)

X X

i-s §

O "2 C

O. 5

k '-3

S 2 8

x a O

£ 2 .£

3 o 3

c Jt c

a T3 60
'£ -

.. -a <u

•- B 60

5 5 |
^ x; S

£ ♦; «

O 3 Z>

DO X 3

60 -t;

E § S

C3 ti tfl

|£ a

'IS.
M ,S _2

•3 C
U

I s. -=

X> "* 60

- -x

^ o **

'I o S

D "> X
Di

2 2

o >

c <u ° 'a

5 a x £

-X . cj

6 - a

Q. so 2
£ .£ E

3 2 c
o 2 «

uca^

o o
2 £

.£• 3 50

S I 5 -

c cS S c

t5 £ & S

*-• ^ > o

fr Tl a

£ u 9 a

5 5 2

_ « 60

■g I -2 E

S x> * 2

o .0 ^ c

o .2

3 B

.2 >»

2 §

■^-< —

2 -S

. c
£ w £
•a 'S

X 60 2

c ~ .£ £

- E -3

o p

.3 n

B S I

S> B 13

.3 "B >

X -3

P £

B. 5
r3 ^>

ISO

i *> s

O j> O

X ^

(D t. O
_> B0 *H

c -5 x

^30

2 - -3

S £

2 3 X
X — ' BO 3

a x ■-

B0 0) <«

3 u S -a
'S ■** * 3

rt o "" ™

y U ^

.H ♦» os 00

C 3 a Pi

•- o •- o

1 £ &|

c^ X Q. *->

visual integration. These attributes have long been discussed in
the literature, but they have never been demonstrated by gross
measurements, either on the brain or on endocasts of modern
Homo sapiens. However, Geschwind and Levitsky (1968) have
shown that when the temporal and parietal lobes are cut away,
the left side shows strikingly enlarged convolutions, the gyri of
Heschl, underneath. Astakhova and Karacheva (1970) have
shown that differences between left and right hemispheres are
present before birth. It is not possible to go into all the
functional details, but they can be taken as a species-specific
attribute of human brain structure, and by extension, of
behavior.

Do the fossil hominid endocasis show such differences?
Unfortunately the endocasts of the South and East African
australopithecines are seldom bilaterally complete, which pre-
cludes any direct measurements. Gross measurements, such as
lengths, arcs, breadths and heights, do not demonstrate any
consistent asymmetries on complete endocasts. LeMay and
Culebras (1972) have suggested that the Neanderthal brain cast
from La Chapelle-aux-Saints shows laterality, but this depends
on how carefully the Sylvian fissure is defined in its posterior
course, a feature generally impossible to observe on most
endocasts. 1 LeMay and Culebras's angioradiography of living
humans does, however, show consistent left-right differences,
but until a more accurate and sensitive method to measure
endocasts of fossil hominids is found cerebral dominance in
fossils cannot be proved. I am currently working on some of the
newer Homo erectus fossils from Indonesia, on the basis of
which a case may be made for cerebral dominance, but it is too
early to be certain. The presence of stone tools, of primitive,
but nevertheless standard patterns, at least 2.6 to 3.0 million
years old, is suggestive of both lateralization and of primitive
communication by a language based on symbols (Holloway,
1969).

'I have not been able to see this fissure clearly on any fossil hominid endocasts I
have examined.

SUMMARY OF DIRECT AND INDIRECT EVIDENCE

In summary, we find among the early australopithecine
examples fairly clear-cut evidence for human, rather than ape-
like, brain organization. This is based on the following evidence:

1. The endocasts show a more human shape, particularly in
the posterior migration of the lunate sulcus, which separates the
primary visual cortex from the parietal association cortex,
signifying an expanded associational cortical zone. The tem-
poral lobe, so often implicated in memory mechanisms, is
expanded in the anterior pole and in the inferior posterior
region. The orbital rostrum is very unlike that of the apes, and
there is a suggestion of an enlargement in the third inferior
frontal convolution, the so-called Broca's area, which is involved
in motor control of speech.

2. Indirectly, the locomotor, manipulatory, dental, and total
skeletal evidence indicates a human musculoskeletal orga-
nization that presumably required neural reorganization to
operate in human behavior patterns.

3. The faunal associations suggest an adaptation based on
scavenging and/or hunting for animal protein. The stone tools
known from this early period are made to standard patterns.
Both the faunal associations and stone tools are indications of
human behavior requiring reorganization at almost all levels of
the brain (Holloway, 1970), from sensorimotor integration and
finesse, through set and attention variables, to memory (the
organization of experience and the storage, recall and reconsti-
tution of elements).

4. Tentative brain :body ratios and encephalization indices
support (but do not prove) a human brain organization.

THE STAGES OF HOMINID BRAIN EVOLUTION:
A POINT OF VIEW

So far I have discussed both direct and indirect evidence to
support the suggestion that the human brain had an early
beginning regardless of its absolute size. All I have said thus far
applies to endocasts, and thus to the brains of the early

hominids. Brains evolve in both material and social contexts. It
is my contention that human social behavior has very old roots,
not only in the sense that we have evolved from some primitive
apelike lineage, but in the sense that human social behavioral
evolution occurred early and was the major stimulus for further
evolution since the time of the australopithecines. I would like
to try to put together the story of the reorganization of the
hominid brain, its great increase in size and the evolution of
human behavior in a synthesis that avoids some of the simplistic
one-to-one linear relationships that physical anthropologists are
prone to make, such as that tools made the brain evolve or that
tools replaced the canines.

In this section I wish to return to the original questions: How
did the human brain evolve to its present state? How can we
interpret the large increase in brain size from Australopithecus
to modern Homo sapiens^.

It is apparent that part of this increase must be related to
increase in body size. Exactly how much is difficult to say,
since it depends on which animal body and brain weights we
compare with man and how we regard "extra" or "vital"
neurons (Jerison, 1963, 1973). Taking the average human brain
weight as 1450 grams and the average body weight as 150
pounds, the following different calculations can be made: (1)
Using Jerison's (1973, p. 44) equation of E = 0.07 P 2/3
(E = brain weight, P = body weight) for higher vertebrates, we
get an expected brain weight of 108 grams for Homo sapiens,
leaving 1342 grams as "extra" (not related to body weight); (2)
If we use Stephan's (1972) equation for "basal insectivores,"
the expected brain weight is 475 grams, leaving 975 grams as
"extra"; (3) Jerison's (1973, p. 391) equation of E = 0.12 P 2/3
for higher primates gives an expected brain weight of 223
grams, leaving 1227 grams as "extra." 1 Both Jerison equations

'I am using "extra" purely in the operational sense that it exceeds a weight based
on a log-log regression with an exponent of roughly 0.6. I do not believe that any
neural elements, are in any other sense "extra," whether in terms of weight or
numbers of neurons. The so-called extras are part and parcel of the animal's adaptive
behavioral repertoire!

leave us with the same degree of encephalization as the dolphin.

Obviously these figures leave much to be desired, as the
formulas are based on regressions relating only to living species.
We would need to know the regressions for our fossil ancestors
(Ramapithecus, Australopithecus, Homo erectus, etc.) to know
what the increase in brain weight relative to body weight has
been. If we use the mean of 442 cc. for the brain weight of the
gracile australopithecines and 45 to 50 pounds as body weight,
the braimbody weight ratio is about 1:45, roughly the same as
modern Homo sapiens. If this ratio remains constant, i.e., at an
exponent of 1.0, then none of the increase (ca. 1000 ml.) is
"extra," at least in terms of the hominid regression equation.

As can be seen, the figures can be used in various ways. It is
all the more curious, then, that, contrary to most opinions
(Jerison, 1963, 1973), the present data on neuron numbers in
the primate cerebral cortex (see, for example, Shariff, 1953) do
not indicate that the increase in brain size in Homo sapiens is
primarily a result of hyperplasia, or the addition of large
numbers of neurons. From Shariff s (1953) data, modern man
seems to have about 1.25 times as many neurons as a healthy
chimpanzee. Jerison's (1963, 1973) calculated "extra" cortical
neurons are at total variance with Shariff s data, the only
empirical evidence existing for primates. According to Jerison
(1963), Homo sapiens has 2.2 times as many cortical neurons as
a chimpanzee, yet his equations for "extra" neurons are derived
from Shariffs empirical histological counts (see Holloway,
1966a, 1974a, for a further critique). 1

From limited neuropathological data, there is a suggestion
that healthy chimpanzees and gorillas might have fewer mature
functioning cortical neurons than human microcephalics (Hollo-
way, 1964, 1968; Lenneberg, 1964, 1967). The behavioral
repertoire of microcephalics is certainly limited, but many of

'This preoccupation on mass can also be found in Count (1973), who transformed
neuron numbers from base 10 to 2; i.e., humans have 2" neurons, while chimpanzees
have 2 31 neurons. Count suggested that thus only two mitotic divisions separates the
chimpanzee from the human brain. I strongly disagree with this interpretation.

them can use language, and their behavior is hardly simian. This
further suggests some basic reorganization of the brain.

Most scientists agree that the major increase in brain size is
most likely related to hypertrophy, or increase in size, of the
elements. The cortical neurons are generally large in man, there
is a reduction in their density and an increase in both dendritic
branching of the receptive processes of the neurons and in the
number of neuroglial cells supporting the neurons. Thus, one
important aspect of the large increase in brain size seems
attributable to the reorganization of numerous component
structures. That is why I believe comparisons based on cranial
capacities alone are meaningless. One cc. of chimp or australo-
pithecine cortex is not equivalent to one cc. of modern human,
Neanderthal or Homo erectus cortex. It is changes in the spatial
relationships between elements that provide our great neural
complexity, for these result in an enormous number of synaptic
contacts, or switching points (Holloway, 1964, 1966b, 1967,
1968).

The great increase in brain size can best be related, I believe,
to a matrix of interacting variables of neural and behavioral
complexity during the Pliocene and Pleistocene epochs that had
an essentially positive feedback structure (see Holloway, 1967).
The matrix involved a change in endocrine-target tissue inter-
action, an increased postnatal dependence of offspring on
parents, delayed maturation and the growing role of social
programming on the brain. This interpretation is based on (1)
observations regarding the effects of hormonal manipulations
on such brain parameters as average cortical neuron size, neuron
density, dendritic branching, glial/neural ratios, and cortically-
mediated behavior; (2) phylogenetic and ontogenetic changes in
cortical histology; and (3) the effects of enriched and deprived
environments on cortical neuron histology. I (Holloway, 1964,
1968) have reviewed this elsewhere and will not repeat the
discussion here. The basic concordance in mammals between
phylogenetic and ontogenetic development and extra environ-
mental training on the one hand, and neurological changes-
decreased neuron density, increased dendritic branching and

increased glial/ncural ratios in animals treated with growth
hormone or thyroxin on the other, is illustrated in table 6. The
table suggests a concordant picture of increase in brain
complexity and cortically-mediated adaptive behavior. Thy-
roidectomy and sensory deprivation, however, produce opposite
results.

TABLE 6

Concordances of Different Lines of Evidence and Various Neural Parameters"

Neural Parameters

Cortically-

Average

Glial:

Amount of

Mediatcd

Size of

Neuron

Neural

Dendritic

Adaptive

Type of Evidence

Neurons

Density

Ratio

Branching

Behavior

Ontogenetic (growth)

Phylogcnetic

(within primates,

related to brain size)

Physiological manipulation

1. throidectomy

2. administration

of thyroxin

3. growth hormone

Environmental manipulation

1. Sensory

deprivation

2. Environmental

complexity and

training augmented

(rats) (ECT vs. IC) 6

a As the size of the neuron increases, so does its perikarya and cytoplasm, thereby
requiring more neuroglial cells to service its metabolic needs. The additional size
means reducing neural density, i.e., the number of neural nuclei in a standard size
cube of cortical tissue. They are thus packed together less tightly. The increased
neuron size also provides more cytoplasmic material for dendritic and axonal
processes. Notice particularly that the hormonal evidence (all of it in vivo) matches
the ontogenetic, phylogenetic, and environmental lines of evidence.

''See Holloway, 1966a, and Rosenzweig, 1972, for details.

Anthropological interpretations of the increase in brain size
generally attempt to relate the increase in cranial capacity
essentially to single aspects of evidence, such as tool-making,

hunting, language, etc. As the fossil hominids show an increase
in endocranial volume, the archaeological record shows a
concomitant increase in the range and sophistication of stone
tool assemblages and in the size and kinds of animals hunted. A
statistical correlation does not, of course, necessarily mean a
causal connection. I find it very difficult, if not impossible, to
draw a causal connection between brain size and stone tools or
hunting habits. These must surely tie in more with social
programming or learning than with an increase in neural
elements.

It would be a great oversimplification, if not a mistake, to
relate cranial capacity in any linear or causal sense to the
increasing complexity of stone tools during the Pleistocene.
Early hominids accomplished more than simply making stone
tools for future archaeologists' digs. Their tools were used in a
variety of different environments, and their cooperative social
behavior was an important part of adaptation to a hunting and
gathering existence. Hunting and associated activities require a
complex organization involving not only perceptual and motor
skills, but an understanding of animals and their habits, plants,
terrain, spoor, tracks, anatomy, butchering techniques, and
perhaps storage. It is the total range of cultural adaptations that
relates to brain increase; the making of stone tools is only one
example, and of course, the most permanently recorded one.

To the extent that the hunting of large animals involved
cooperative enterprise, selection would certainly have favored
behavioral mechanisms facilitating communication, including
symbolic language. Language would have led to increased
complexity of social interaction, involving appreciation of
numerous related cues from social and material environments,
and the control and inhibition of responses. In short, the
increasing complexity of stone tools indicates other processes,
but it cannot lead to more than educated guesses about the
ecological complexity of selection pressures for human biosocial
adaptations. (These relationships between tool-making and
language, and hunting behavior and various levels of neural

structure have been examined in greater detail by Holloway
1964, 1969, 1970.)

Although the australopithecine brains were small, they were
larger, both relatively and absolutely, than those of the
chimpanzees, which probably had similar body weights. Be-
tween the chimpanzee or the gorilla and man there is a large
difference in the duration of the growth period. Maturation is
complete in a chimpanzee at nine to 1 1 years, whereas in man it
takes about 20 to 25 years.

As yet we cannot look at a fossil and say at what age it
became fully adult; but we must assume that growth rates and
durations changed over the course of human evolution. One
cannot get a brain to evolve in size without prolonging the
period of its growth. Growth is a complex process involving
interaction among genetic instructions for locus and timing,
tissue differentiation, hormone environment (growth hormones,
thyroxin, and androgens) and proper nourishment (including
social nourishment). One of the organs most vulnerable to
malnourishment is the growing brain, particularly during
periods of mitotic division and nerve cell enlargement. The
earliest evidence of increase in brain size in the fossil record
coincides with the earliest evidence for utilization of protein-
rich food (animal flesh). It seems an inescapable conclusion that
there was an adaptive relationship between hunting and the
evolution of the brain, mediated through longer periods of
growth and dependence.

A SPECULATIVE MODEL OF HOMINID EVOLUTION

What follows is a set of speculations concerning the interrela-
tions among a number of complex variables at different levels
(anatomical, physiological, neuroanatomical, ecological, and
social). The main purpose of this model is merely to show the
matrix of variables that I believe must be considered if we are to
have a clearer understanding of how the human brain evolved.

Beginning with Ramapithecus (10 to 14 million years ago)

one can postulate that adaptations based on a savanna environ-
ment (utilization of seeds, grass, and other vegetation) led to
strong positive selection for bipedalism. I do not think we can
speculate further without additional material. Consequently,
my model starts after the Ramapithecus level of adaptation.

Stage 1 : Early australopithecine phase. Major emphasis on social behav-
ior adaptations, involving bipedalism, endocrine organization, and
brain reorganization.

Stage 2: Late australopithecine-"habiline" phase. Major emphasis on
consolidation and refinement of Stage 1 .

Stage 3: Late "habiline"-early Homo erectus to Neanderthal-sapiens
phase. Emphasis on elaboration of cultural skills through a positive
feedback relationship and brain enlargement.

Stage 1 includes the rudimentary development of coopera-
tive, sex-role-separated social groups resulting from endocrine
changes involving hormones and target-tissues. There was a
reduction of sexual dimorphism in tooth and skeletal size and
an increase in epigamic features of secondary sexual characteris-
tics such as permanent breasts and fat distribution. There were
possibly other changes facilitating continuous sexual receptivity
of the female and closer affective relations between the sexes.
This complex of correlated anatomical, physiological, and
behavioral changes led to greater sexual and social control
associated with prolonged periods of postnatal dependence and
learning. Changes in the interactions between hormones and
target-tissues could have led to a reduction in aggressive
components of behavior, sexual dimorphism in size and
increased periods of growth with delayed maturation of skeletal
development. These processes are mediated in a complex
manner by the androgens and involve other hormones as well.
The endocrine changes that led to the dimorphic features cited
above could have played an important role in decreasing
intragroup aggression, permitting groups to live more densely.
In other words, the changes led to an increase in cooperative
behavior (both among males and females and among males) that
meant a stronger protection against both predators and other
hominid groups. At the same time they affected growth rates,

accounting for longer periods of dependency and postnatal
growth during which the brain showed an allometric increase.
Associated with this complex of correlated changes are the
developments of language (using a primitive symbol system) and
hunting and scavenging (with a greater effective range due to
more advanced bipedal locomotion).

I regard the development of language as more closely bound
up with social affect and control than with hunting behavior
involving signaling and "object naming," although this does not
mean that hunting could not have been a strong positive
selection factor for language. 1 In addition to reorganization of
social behavior and bipedal adaptation, there was a reorganiza-
tion of the brain involving, minimally, a decrease in primary
visual cortex on the convex cerebral surface and an increase in
parietal and temporal association cortex, allowing for greater
discrimination among complex cues of the environment and for
extension of foresight and memory to cope more effectively
with the savanna-type environment. Associated with these is the
early manufacture of stone tools to extend the economic base.
The tools may have been used to break bones to secure marrow
and to detach peices of flesh or skin. They may also have been
used as missiles to drive off carnivores from their kills. The
latter behavior involved not only cooperation among group
members, but skill in coordinating hand-eye movements and a
complex appreciation of spatial-visual calculation. It is very
tempting to relate this kind of behavior with the right
hemisphere, known to be dominant in such coordination.

Stage 2 includes refinement and elaboration of the changes in
social behavior begun in Stage 1, as well as an increased
dependence on social cohesion, language, and stone tools.

'I do not agree that human cognition, and more particularly spoken and gestural
communication are mainly cortical-to-cortical events, "liberated," so to speak, from
limbic influences. Emotional involvement and tonus is always present in human
communication, except perhaps in cases of psychopathology. This does not mean
that evolutionary changes in cortical tissue and hemispheric relationships were not
necessary. I only mean that those changes were not merely additive, but totally
integrated with noncortical structures, particularly the thalamus, limbic structures,
hippocampus, and reticular formation.

Bipedal locomotion was essentially fully human. There was
both relative and absolute expansion of the brain, associated
mainly with increased body size. There was greater efficiency of
economic sharing and cooperation between the sexes, 1 provid-
ing the basis for longer periods of postnatal dependency and
learning, which initiated a feedback system between brain and
cultural behavior. Language behavior became more strongly
developed, and cognitive behavior of a more nearly human type
developed, where language and tool-making arose from the same
psychological structuring. There were true stone tool cultures at
this stage, and language had prime importance in maintaining
social cohesion and control and in "programming" offspring.
Dependence on hunting increased and there was more success in
stalking and hunting larger game. There was a selection for
increased body size, bipedal agility and predictive abilities for
more successful hunting. The social behavioral changes outlined
in Stages 1 and 2 permitted longer male-male association for
persistent hunting and for the protection of a more secure home
base for females and young, who were providing small game and
vegetables. The "initial kick," or "human revolution," is fully
set and leads to Stage 3.

In Stage 3 a positive feedback between brain development
and cultural complexity was mediated through the increased
periods of dependency and learning (which was taking place in a
more complex and stimulating material and social environment)
of the offspring. The major neural changes are those of size and
refinement of the reorganized human brain (that is, sensori-
motor, associative, extrapyramidal modulation, and cerebellar
involvement in manual dexterity). This is not a stage of
behavioral innovation, but an elaboration of "complexity-
management" involving fineness of sensory discrimination and
association between larger sets of past memories and skills (see
Holloway, 1967).

'No chauvinistic intents are harbored in the speculative model, in terms of either
male or female superiority. I view the evolution of sex differences, both in behavior
and morphology as complementary to human evolution, not as competitive or
supraordinative.

It must be emphasized that I see these stages as gradual and
continuous, with certain developments stressed more strongly in
one stage than in another. My main point is to show that social
behavior mechanisms have had a long development, beginning
with the early hominids. In a sense, increase in brain size is
minor compared to the evolution of the social matrix. Brain
expansion finally depends on a solid behavioral foundation. My
model takes into account both the skeletal remains and the
cultural evidence and provides a base for synthesizing anatomi-
cal, behavioral (social and individual), physiological, adapta-
tional, and ecological variables.

It is possible, I believe, to consider more molecular analyses
within this model. At the level of neuroanatomy, one can
suggest various brain regions that could be correlated with
behavioral attributes such as set and attention, concentration,
"memory" (permanence, quantity, facility, and strategy of
recall), hand-eye and running coordination, mother-infant af-
fect, babbling and reticular core reorganization, cerebral laterali-
zation, play, curiosity, prolongation of prepubertal vividness of
experience, memory, and so on. To do so, however, is far
beyond the limits of this lecture.

It must be understood that the analysis of endocranial casts
alone cannot play more than a limited role in elaborating my
hypothesis, or in supporting my speculations. The external
morphology of endocasts provides clues, not proof, about past
selection pressures, and these clues are fairly gross. The
judicious use of endocasts, both as clues to neural reorganiza-
tion and to changes of growth variables must await further
discoveries with firm dates. While studies of australopithecine
endocasts are in progress, it should be apparent that the
specimens have potential use, both as clues to general events in
hominid evolution and as morphological patterns for taxonomic
purposes. The analysis given thus far shows, I believe, that the
evolution of the brain has always been an integral part of
hominid evolution and was not something that took place
following other changes in different morphological sectors of
the hominids.

Let me close by asserting my belief that human behavior is a
long-standing evolutionary development, possibly more than
three million years old. Human thought, aside from its more
sophisticated scientific recency, is no late invention, but instead
is very old. The human brain is both the product and cause of
the evolution of human social behavior, and we should
recognize that our brains are both the instruments and products
of our sociality, the genesis of which was long in the making.

ACKNOWLEDGMENTS

I am grateful to Drs. Ian Tattersall and Lester Aronson and
the American Museum of Natural History for honoring me with
this lecture and for helping me with its publication. Much of the
actual work on fossil specimens was carried out in Africa,
Europe, and Indonesia with support from the National Science
Foundation. Without the kindness and cooperation of Dr. P. V.
Tobias at the University of Witwatersrand, South Africa, and
Drs. Louis and Mary Leakey and Mr. Richard Leakey of the
National Museum in Nairobi, Kenya, this work would not have
been possible, and I am indebted to them and their staff for
their help. I am similarly grateful to Dr. Teuku Jacob, Gadjah
Mada University, Indonesia, and his staff for their courtesy,
help, and hospitality; to Dr. Darwin Kadar, Section of Palaeon-
tology, Geological Institute, Bandung, Indonesia; to Dr. G. H.
R. von Koenigswald, Senckenberg Institute, Frankfurt,
Germany. I owe special thanks to Dr. Alan Walker, now at
Harvard University, for his cooperation, help, facilities, and
friendship. Almost all the chimpanzee and gorilla specimens
endocasted and studied were lent to me by the American
Museum of Natural History, through the courtesy of Drs. Ian
Tattersall and Richard Van Gelder, to whom I am most grateful.
I thank Dr. Jeffrey Schwartz and Mr. Barry Cerf of Columbia
University for their help in keeping me supplied with specimens,
and Ms. Kate McFeely for typing the original manuscript. This
research has also been supported by a Guggenheim Fellowship,
for which the author is most grateful.

LITERATURE CITED

Astakhova, A. T., and A. A. Karacheva
1970. Asymmetry of the brain in the human fetus. Trans. Krasnoyarsk Med. Inst.,
vol.9 (5), pp. 9-12. (Translated from Ref. Zhur. Biol., no. 4M432, 1971.)

Bailey, P., G. von Bonin, H. W. Carol, and W. S. McCullough
1943. Long association fibres in cerebral hemisphere of monkey and chimpanzee.
Jour. Neurophysiol., vol. 6, pp. 129-134.

Clark, W.E.LeGros

1947. Observations on the anatomy of the fossil Australopithecinae. Jour. Anat.,
vol. 81, pp. 300-334.

Cobb.S.
1965. Brain size. Arch. Neurol., vol. 12, pp. 555-561 .

Count, E. W.
1973. Being and becoming human. New York, D. Van Nostrand Co.

Dart, R.
1925. Australopithecus africanus: the man-ape of South Africa. Nature, vol. 115,
pp. 195-199.

Edinger, T.
1929. Die fossilen Gehirne. Ergeb. Anat. Entwicklungsgesch., vol. 28, pp. 1-249.
1949. Paleoneurology versus comparative brain anatomy. Confin. Neurol., vol. 11,

pp. 5-24.
1964. Recent advances in paleoneurology. Progr. Brain Res., vol. 6, pp. 117-160.

Geschwind, N., and W. Levitsky

1968. Human brain: left -right asymmetries in temporal speech region. Science,
vol. 161, pp. 186-187.

Hirschler, P.
1942. Anthropoid and human endocranial casts. Ph.D. thesis, Amsterdam, N. V.
Noord-Holiandsche Uitgeber Maat Schappi.

Holloway, R. L.
1964. Some aspects of quantitative relations in the primate brain. Ph.D.

dissertation, Univ. California, Berkeley.
1966a. Cranial capacity and neuron number: a critique and proposal. Amer. Jour.

Phys. Anthrop., vol. 25, pp. 305-314.
1966b. Dendritic branching: some preliminary results of training and complexity in

rat visual cortex. Brain Res., vol. 2, pp. 393-396.

1967. The evolution of the human brain: some notes toward a synthesis between
neural structure and the evolution of complex behavior. Gen. Systems, vol.
12, pp. 3-19.

1968. The evolution of the primate brain: some aspects of quantitative relations.
Brain Res., vol. 7, pp. 121-172.

1969. Culture: a human domain. Current Anthrop., vol. 10, pp. 395-412.

1970. Neural parameters, hunting and the evolution of the primate brain. In
Noback, C. R., and W. Montagna (eds.). The primate brain. New York,
Appleton-Century-Crofts, pp. 299-310.

1972a. New australopithecine endocasts, SK 1585, from Swartkrans, South Africa.

Amer. Jour. Phys. Anthrop., vol. 37, pp. 173-186.
1972b. Australopithecine endocasts, brain evolution in the Hominoidea, and a

model of human evolution. In Tuttle, R. (ed.), Functional and evolutionary

biology of the primates. Chicago, Aldine, pp. 185-204.
1974a. On the meaning of brain size: review of H. J. Jerison's evolution of the

brain and intelligence (1973). Science, vol. 184, pp. 677-679.
1974b. The casts of fossil hominid brains. Sci. Amer., vol. 231, pp. 106-115.

Jerison, H. J.

1963. Interpreting the evolution of the brain. Human Biol., vol. 35, pp. 263-291.
1973. Evolution of the brain and intelligence. New York, Academic Press.

Keith, A.
1931. New discoveries relating to the antiquity of man. London, Williams and
Norgate, Ltd.

LeMay, M., and A. Culebras
1972. Human brain: Morphological differences in the hemispheres demonstrated
by carotid arteriography. New England Jour. Med., vol. 287, pp. 168-170.

Lenneberg, E. H.

1964. A biological perspective of language. In Lenneberg, E. H. (ed.), New
directions in the study of languages. Cambridge, Massachusetts Institute of
Technology Press.

1967. Biological foundations of language. New York, John Wiley and Sons, Inc.

Lovejoy, C. O., and K. G. Heiple
1970. A reconstruction of the femur of Australopithecus africanus. Amer. Jour.
Phys. Anthrop., vol. 32, pp. 33-40.

MacKinnon, I. L., J. A. Kennedy, and T. V. Davies
1956. The estimation of skull capacity from roentgenologic measurements. Amer.
Jour. Roentgenology, Radiation Ther., Nucl. Med., vol. 76, pp. 303-310.

Pakkenberg, H., and J. Voigt
1964. Brain weight of the Danes. Acta Anat., vol. 56, pp. 297-307.

Pilbeam, D. R.
1969. Newly recognized mandible of Ramapithecus. Nature, vol. 222, pp. 1093.

Radinsky, L.
1967. The oldest primate endocast. Amer. Jour. Phys. Anthrop., vol. 27, pp.
385-388.

1970. The fossil evidence of prosimian brain evolution. In Noback,C. R.,and \V.
Montagna (eds.). The primate brain. New York, Appleton-Century-Crofts,
pp. 209-224.

Rosenzweig, M. R.
1972. Brain changes in response to experience. Sci. Amer., vol. 226, pp. 22-29.

Schepers, G.W. H.
1946. The endocranial casts of the South African ape-men. Transvaal Mus. Mem.,
no. 2, pp. 155-272.

Shariff, M.
1953. Cell counts in the primate cerebral cortex. Jour. Comp. Neurol., vol. 98, pp.
38M00.

Simons, E. L.
1961. The phyletic position of Ramapithecus. Postilla, no. 57.
1964. On the mandible of Ramapithecus. Proc. Nat. Acad. Sci., vol. 51 , pp. 528.
1969. Late Miocene hominid from Fort Ternan, Kenya. Nature, vol. 221 , pp. 448.

Stephan, H.
1972. Evolution of primate brains: a comparative anatomical investigation. In
Tuttle, R. (ed.), The functional and evolutionary biology of primates.
Chicago, Aldine-Atherton, pp. 155-174.

Tobias, P. V.
1967. Olduvai Gorge. Vol. II. Cambridge, Cambridge Univ. Press.

1971 . The brain in hominid evolution. New York, Columbia Univ. Press.

r;2bi A

J35

0.44

974

FORTY-FOURTH
]jAMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

1974

PERSISTENT PROBLEMS IN

THE PHYSICAL CONTROL

OF THE BRAIN

ELLIOT S. VALENSTEIN

<? *•

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1975

lit

1869
THE LIBRARY

FORTY-FOURTH

JAMES ARTHUR LECTURE ON

THE EVOLUTION OF THE HUMAN BRAIN

FORTY-FOURTH
/JAMES ARTHUR LECTURE ON
THE I VOLUTION OF THE HUMAN BRAIN

1 9 74

PERSISTENT PROBLEMS IN THE
PHYSICAL CONTROL OF THE BRAIN

ELLIOT S. VALENSTEIN

Professor of Psychology and Neurosciencc
University of Michigan, Ann Arbor

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1975

JAMES ARTHUR LECTURES ON
Till EVOLUTION OF 1111 HUMAN BRAIN

Frederick I ilney . I'h< Brain in Relation to Behavior; March 15, 1932

C. Unison Herrick, Brains as Instruments of Biological Values; April 6, 1933

D. M. S. Watson, The Story o) Fossil Brains front Fish to Man; April 24, 1934

C. U. Ariens Kappcrs. Structural Principles in the Nervous System, The Develop-
ment of the Forebrain in Animals and Prehistoric Hitman Races, April 25, 1935

Samuel T. Orton, Tlie Language Area of the Human Brain and Some of its
Disorders; May 15, 1936

R. W. Gerard, Dynamic Neural Patterns; April 15, 1 937

Franz Weidenrcich, The Phylogenctic Development of the Hominid Brain and its
Connection with the Transformation of the Skull; May 5, 1938

G. Kingsley Noble, The Neural Basis of Social Behavior of Vertebrates; May 11,
1939

John F. Fulton, A Functional Approach to the Evolution of the Primate Brain; May
2, 1940.

Frank A. Beach, Central Nervous Mechanisms Involved in the Reproductive
Behavior of Vertebrates; May 8, 1 94 1

George Pinkley, A History of the Human Brain; May 14, 1942

James W. Papez, Ancient Landmarks of the Human Brain and Their Origin; May 27,
1943

James Howard McGregor, The Brain of Primates; May 1 1, 1944

K. S. Lashlcy , Neural Correlates of Intellect; April 30, 1945

Warren S. McCulloch, Finality and Form in Nervous Activity; May 2, 1946

S. R. Detwilcr, Structure-Function Correlations in the Developing Nervous System
as Studied by Experimental Methods; May 8, 1947

Tilly Fdinger, The Evolution of the Brain; May 20, 1948

Donald O. Hcbb, Evolution of Thought and Emotion; April 20, 1949

Ward Campbell Halstead, Brain and Intelligence; April 26, 1950

Harry F. Harlow, The Brain and Learned Behavior ; May 10, 1951

Clinton N. Woolsey, Sensory and Motor Systems of the Cerebral Cortex; May 7,
1952

Alfred S. Romer, Brain Evolution in the Light of Vertebrate History; May 21 , 1953
Horace W. Magoun, Regulatory Functions of the Brain Stem; May 5. 1954

*Fred A. Mettler. Culture and the Structural Evolution of the Neural System; April
21, 1955

*Pinckney J. Harman, Paleoneurologic, Neoneurologic, and Ontogenetic Aspects of
Brain Phytogeny, April 26, 1956

*Davenport Hooker, Evidence of Prenatal Function of the Central Nervous System in
Man; April 25, 1957

*David P. C. Lloyd, The Discrete and the Diffuse in Nervous Action; May 8, 1958

*Charles R. Noback, The Heritage of the Human Brain; May 6, 1959

*Ernst Scharrer, Brain Function and the Evolution of Cerebral Vascularization; May
26, 1960

Paul I. Yakovlev, Brain, Body and Behavior. Stereodynamic Organization of the
Brain and of the Motility-Experience in Man Envisaged as a Biological Action
System; May 16, 1961

H. K. Hartline, Principles of Neural Interaction in the Retina; May 29, 1962

Harry Grundfest, Specialization and Evolution of Bioelectric Activity; May 28,
1963

*Roger W. Sperry , Problems Outstanding in the Evolution of Brain Function; June 3,
1964

*Jose M. R. Delgado, Evolution of Physical Control of the Brain; May 6, 1965

Seymour S. Kety, Adaptive Functions and the Biochemistry of the Brain; May 19,
1966

Dominick P. Purpura, Ontogenesis of Neuronal Organizations in the Mammalian
Brain; May 25, 1967

*Kenneth D. Roeder, Three Views of the Nervous System; April 2, 1968

t Phillip V. Tobias, Some Aspects of the Fossil Evidence on the Evolution of the
Hominid Brain; April 2, 1969

*Karl H. Pribram, What Makes Man Human; April 23, 1970

Walle J. H. Nauta, A New View of the Evolution of the Cerebral Cortex of Mammals;
May 5. 1971

David H. Hubel, Organization of the Monkey Visual Cortex; May 1 1, 1972

Janos Szentagothai, The World of Nerve Nets; January 16, 1973

Ralph L. Holloway, The Role of Human Social Behavior in the Evolution of the
Brain; May 1, 1973

* Published versions of these lectures can be obtained from The American Museum of
Natural History, Central Park West at 79th St., New York, N.Y. 10024.

-("Published version: The Brain in Hominid Evolution, New York: Columbia University
Press, 1971.

PERSISTENT PROBLEMS IN THE PHYSICAL
CONTROL OF THE BRAIN 1

INTELLECTUAL AND SOCIAL CLIMATE
AND SCIENTIFIC DISCOVERY

There is great temptation to dramatize scientific discoveries
by picturing them as the result of sudden insights or lucky
accidents. In actuality, this is seldom the entire story since most
discoveries also reflect the intellectual and social climate of
their time. The many potentially significant observations that
were neglected or misinterpreted attest to the importance of a
prepared mind. It is true, for example, that Luigi Galvani
observed a suspended frog twitch in synchrony with flashes of
lightning, but this event was significant because it occurred at a
time of increasing interest in the relation of physical and
biological phenomena. Galvani's observation occurred only a
short time after Mesmer's suggestion that "animal magnetism"
was the basis of what is now called hypnotism. Not very many
years earlier, the Scottish anatomist John Hunter and the
English physicist Henry Cavendish speculated that the study of
such electric fish as the eel and torpedo (an electric ray fish)
might help to explain the action of nerves in general. In fact,
the part played by electric fish in the early history of
bioelectricity and electrotherapy has been the subject of an
interesting essay by Kellaway (1946).

Galvani's observations, therefore, were not entirely acci-
dental. It is certain that Galvani did not suspend a frog between
a wire attached to a lightning rod and a rod immersed in a well
by mere chance. He was very much aware of Benjamin
Franklin's demonstration that atmospheric electricity could be

'A comprehensive discussion of the historical, scientific, and ethical considera-
tions related to the physical control of the brain was presented by Valenstein (1973).
The research reported in the present paper was supported by NIMH Research Grant
2 ROl MH20811-03.

tapped in a harmless manner. This particular frog experiment
was clearly only one of many Galvani designed to study the role
of electricity in biological phenomena. Many of these experi-
ments involved the observation that a frog's muscle would
twitch when touched with metal probes. The lively dispute
between Galvani and the physicist Allesandro Volta that took
place between 1790 and 1798 was over interpretation. Galvani
argued for the existence of "animal electricity," whereas Volta
argued for "metallic electricity" and claimed that the dissimilar
metals used in most of Galvani's experiments produced an
electric force that caused the muscles to contract. This
controversy shaped much of the research on the nervous system
during the early part of the nineteenth century. By 1848 when
du Bois-Reymond published his book, Investigations of Animal
Electricity, and Helmholtz had shown that the speed of nerve
conduction was very different from that of electric current, the
controversy had disappeared.

The value of electrical stimulation to study the nervous
system, however, increased in importance with the passage of
time. The technique, which had been applied primarily to the
crural nerve and gastrocnemius muscle of the frog, began to be
applied directly to mammal brains. Legend has it that while
dressing the head wounds of soldiers, Eduard Hitzig observed
that their muscles twitched on the side of the body opposite the
injury. The 1870 report by Fritsch and Hitzig describing the
frontal lobe regions of dogs from which electrical stimulation
could evoke bodily movement is traditionally attributed to this
accidental observation. As Doty pointed out, there is no truth
in this legend despite the number of writers who delight in
repeating it. Fritsch and Hitzig had become embroiled in the
controversy over specific versus holistic representation of func-
tions within the brain, particularly the cerebral cortex. Many in-
vestigators were using electrical stimulation to settle the issue but
the results were often confusing because it was not yet appreciated
that Galvanic (direct) current destroyed nerve tissue. (Du Bois-
Reymond had already developed an inductorium for providing
alternating or faradic current, but it was not universally used.)

Fritsch and Hitzig concluded that their results showed clearly
that "some pyschological functions and perhaps all of them . . .
need certain circumspect centers of the cortex." David Ferrier
reached the same conclusion a few years later as a result of his
electrical stimulation of the monkey cortex. Friedrich Goltz, on
the other hand, argued for holism by describing dogs that were
still capable of moving all their limbs after removal of virtually
half the brain. The literature of the period provides much
support for the statement attributed to Alfred Binet, "Tell me
what you are looking for and I'll tell you what you will find."

Even the first known attempt at psychosurgery must be
examined against the background of the localization contro-
versy. In 1891, Gottlieb Burckhardt, the director of the insane
asylum at Prefargier, Switzerland, reported the results of
removing part of the cortex of six "demented" patients. He
said: "Who sees in psychoses only diffuse illness of the
cortex . . . for him it will naturally be useless to remove small
parts of the cortex in the hope to influence a psychosis
beneficially by this means. One has to be as I am, of a different
opinion. That is, our psychological existence is composed of
single elements, which are localized in separate areas of the
brain .... Based on these considerations and theories expressed
earlier, I believe one has the right to excise such parts of the
cortex, which one can consider starting points and centers of
psychological malfunctions and furthermore, to interrupt con-
nections whose existence is an important part of pathological
processes."

Controversies about localization are still with us, but of
course at a more sophisticated level. Stereotaxic techniques and
reliable methods for permanently implanting electrodes have
made it possible to undertake behavioral (psychological) studies
over long periods of time. Earlier controversy was about simple
motor responses; current arguments often focus on the localiza-
tion of relatively complex motivational states. The intellectual
and social climate also influences contemporary research, but it
is difficult to achieve adequate perspective when one is very
close to a scene. Nevertheless, it is helpful to try. I believe I can

discern two major influences that have shaped brain stimulation
studies from 1950 to the present. One of these involves the
attempt to accumulate evidence demonstrating that electrical
stimulation of discrete subcortical brain areas can evoke natural
drive states. The other influence, which stems directly from the
first, has been the preoccupation with brain stimulation as a
technique for controlling behavior.

For psychologists interested in studying the process of
learning, the early 1950s was a time of increasing disillusion-
ment with theories based on changes in hypothetical drive states
assumed to take place in the brain. (Indeed, this was a period
when it was often maintained that CNS, the common abbrevia-
tion for the central nervous system, in reality meant the
"conceptual nervous system.") These drive-reduction learning
theories, as they are called, emphasized that we learn only (or in
the weaker versions of the theory, we learn best) those
stimulus-response connections that are associated with changes
in level of drive state. Although it was recognized that
peripheral body factors may contribute to drive state, a number
of experiments had made it evident that drives such as hunger
and thirst did not depend upon intensity of stomach contrac-
tions, dryness of mouth, or other obvious bodily cues. Drive
states, therefore, were presumed to be represented mainly by
the level of activity in functionally specific neural systems
within the brain. However, this conclusion was inferential, and
therefore the properties of drive, the major variable in the
theory, had to be inferred and could not be measured. The field
was rapidly degenerating into unresolvable arguments of little
interest to anyone not indoctrinated into this specialty.

Drive-reduction theorists desperately needed some new input
into their system. Although the Swiss physiologist Walter Hess
had received a Nobel prize by this time, the details of his
German publications were not well known in the United States.
Hess had been stimulating the diencephalon in cats, using a
technique that permitted him to study the responses evoked in
awake, relatively unrestrained animals. Most of his observations
were directed toward understanding the regulation of so-called

autonomic responses such as changes in pupil size, blood
pressure, heart rate, respiration, and the like. When Hess was
invited to speak at Harvard in 1952, a number of people became
aware tor the first time that some of his studies seemed to
demonstrate that electrical stimulation of certain areas in the
diencephalon could suddenly make peaceful cats aggressive or
satiated cats hungry. These reports were seized upon, for they
seemed to provide a means to manipulate drives and to measure
them directly.

Neal Miller (1973, pp. 54-55) reflected on his initial interest
in brain stimulation studies and described it as follows:

'if 1 could find an area of the brain where electrical
stimulation has the other properties of normal hunger, would
the sudden termination of that stimulation function as a
reward? If I could find such an area, perhaps recording from it
would provide a way of measuring hunger which would allow
me to see the effects of a small nibble of food that is large
enough to serve as a reward, but not large enough to produce
complete satiation. Would such a nibble produce a prompt,
appreciable reduction in hunger, as demanded by the drive-
reduction hypothesis?"

This certainly does not reflect the sophistication of Miller's
current thinking on the problem, but it does illustrate the
earlier intellectual climate that produced a need to find
similarities between such behaviors as eating, drinking, and
aggression when elicited by brain stimulation, and the same
behaviors when motivated by natural internal states. What was
found was that eating, drinking, grooming, gnawing, aggression,
foot-thumping, copulation, carrying of young, and many other
behaviors could be triggered by brain stimulation. What was
claimed was that discrete brain centers were identified which,
when stimulated electrically, would evoke specific and natural
states such as hunger, thirst, sexual appetite, and maternal
drives. Tests were designed to emphasize the naturalness of the
evoked states and dissimilarities were disregarded or dismissed
as experimental noise. A personal experience illustrates the
influence of the prevailing bias. When reporting at a meeting

that the same brain stimulus frequently evoked eating, drinking,
and other behaviors, I noted that these and other observations
raised some serious questions about the belief that natural drive
states were evoked. A colleague attending the session told
me that he had made similar observations several years
earlier, but as they interfered with the planned experiments,
the testing conditions were arranged so that the stimu-
lated animals had no chance to express these "irrelevant"
behaviors.

In addition to overlooking behavioral observations inconsis-
tent with the assumption that natural drive states could be
duplicated by stimulating single points in the brain, several
other trends characterized the period from 1955 to 1970. There
was a tendency to rush into print with every new observation of
a different behavior that could be evoked by brain stimulation.
The competition for priority of discovery and the need to
demonstrate progress to the granting agencies often interfered
with any serious attempt to understand the relation between
brain stimulation and behavior change. One active researcher
remarked to me that he would not be "scooped" again,
bemoaning the fact that someone had published an article
describing a new behavior that could be evoked by brain
stimulation before he had. The list of such behaviors kept
growing. One other factor that had a major impact was the
belief that each evoked behavior was triggered from different
and discrete brain sites. In some cases, reports encouraging the
growth of this belief actually presented no anatomical informa-
tion, but despite this deficiency, there was little hesitancy in
using loosely defined anatomical terms (really pseudoanatomi-
cal) such as the "perifornical drinking area." Some reports
presented very complete histological data, but where the
authors emphasized the separateness of brain areas eliciting
different behaviors, others with a different bias could just as
readily see diffuse localization and considerable overlap. In
total, the impression was created that a large number of natural
motivational states could be reliably controlled by "tapping
into" discrete brain sites.

POPULARIZATION OF RESEARCH

As the reports of these experiments began to be dissemi-
nated, a number of other distortions were introduced. These
accounts fed the growing fear that this new brain technology
might be used to control human behavior. The emphasis on
control, by numerous demonstrations of behavior being turned
"on and off" and by selective and oversimplified descriptions of
these demonstrations in the popular press, has had the
predictable effect. The possibility of behavior control by
various brain interventions has become a popular topic for
novels, television shows, movies, magazines, feature articles in
newspapers, and even essays purporting to describe life in the
not too distant future. Michael Crichton's The Terminal Man is
only one of many novels that have used this theme. It may be
no exaggeration to say that this story may have a greater impact
(because it is believed by more people) than Mary Shelley's
Frankenstein. Taking a different tack, an article that appeared
in Esquire magazine described a government of the future, an
"electroligarchy," where everyone is controlled by electrodes
(Rorvik, 1969). It is not necessary to demonstrate that all this
material is believed by everyone, or even by most people, in
order to recognize that the virtual bombardment from the
media has had a profound effect.

Even the material meant only for amusement, and not
intended to be taken seriously, gradually begins to become a
part of our serious thinking and influences our perception of
interpersonal relations. A New York Times article dated
September 12, 1971, described the scientists who: "have been
learning to tinker with the brains of animals and men and to
manipulate their thoughts and behaviors. Though their methods
are still crude and not always predictable, there can remain little
doubt that the next few years will bring a frightening array of
refined techniques for making human beings act according to
the will of the psychotechnologist.' ,

With more drama and expressing less reservation, Perry
London (1969, p. 37) a professional psychologist, stated that

All the ancient dreams of mastery over man and all the tales of zombies,
golems, and Frankensteins involved some magic formula, or ritual, or
incantation that would magically yield the key to dominion. But no
one could be sure, from the old Greeks down to Mrs. Shelley, either by
speculation or vivisection, whether there was any door for which to
find that key . . . This has been changing gradually, as knowledge of the
brain has grown and been compounded since the nineteenth century,
until today a whole technology exists for physically penetrating and
controlling the brain's own mechanisms of control. It is sometimes
called "brain implantation," which means placing electrical or chemical
stimulating devices in strategic brain tissues . . . These methods have
been used experimentally on myriad aspects of animal behavior, and
clinically on a growing number of people . . . The number of activities
connected to specific places and processes in the brain and aroused,
excited, augmented, inhibited, or suppressed at will by stimulation of
the proper site is simply huge. Animals and men can be oriented toward
each other with emotions ranging from stark terror or morbidity to
passionate affection and sexual desire . . . Eating, drinking, sleeping,
moving of bowels or limbs or organs of sensation gracefully or in spastic
comedy, can all be managed on electrical demand by puppeteers whose
flawless strings are pulled from miles away by the unseen call of radio
and whose puppets made of flesh and blood, look "like electronic
toys," so little self-direction do they seem to have.

It is little wonder that the feeling of being controlled by
surreptitiously implanted brain devices has become an increas-
ingly common delusion in paranoia.

While many people emphasize the potential misuse of these
new brain-manipulating techniques, there are some who have
stressed what they believe is their positive potential. They see in
them a possible cure not only for intractable psychiatric
disorders, but for intractable social problems as well— particu-
larly those related to violent crimes and wars. This potential of
brain intervention to achieve desirable ends has been expressed
by Kenneth Clark in his presidential address to the 1971
convention of the American Psychological Association: Clark
suggested that "we might be on the threshold of that type of
scientific biochemical intervention which could stabilize and
make dominant the moral and ethical propensities of man and
subordinate, if not eliminate, his negative and primitive behav-
ioral tendencies."

Proposals of this type can best be discussed after a more

realistic foundation is prepared for critically examining the
capacity of physical techniques to modify brain-behavior
relationships.

A CRITICAL EXAMINATION OF THE EVIDENCE

It should be recognized from the outset that evidence limited
to the demonstration of inhibition or evocation of some
behavior pattern can be very misleading. Such demonstrations
convey the impression that there is a simple and predictable
relationship between specific brain sites and complex behavior
patterns. Also, the implication that only one behavior is
influenced by the electrical stimulation encourages the infer-
ence that the control is very precise and selective.

It might not be inappropriate to begin the critical examina-
tion with a demonstration that is familiar to most people,
Delgado's (1969) purported demonstration of brain stimulation
inhibiting aggressiveness in a bull. An article in the New York
Times (September 12. 1971) described the event as it is
typically reported: "Dr. Delgado implanted a radio-controlled
electrode deep within the brain of a brave bull, a variety bred to
respond with a raging charge when it sees any human being. But
when Dr. Delgado pressed a button on a transmitter, sending a
signal to a battery-powered receiver attached to the bull's horns,
an impulse went into the bull's brain and the animal would
cease his charge. After several stimulations, the bull's naturally
aggressive behavior disappeared. It was as placid as Ferdinand."

Although this interpretation is commonly accepted, there is
actually little evidence supporting the conclusion that the
stimulation had a specific effect on the bull's aggressive
tendencies. A viewing of the film record of this demonstration
should make it apparent to all but the most uncritical observer
that the stimulation forced the bull to turn in circles in a very
stereotyped fashion. This should not surprise anyone familiar
with the brain, as the stimulating electrode was situated in the
caudate nucleus, a structure known to play an important role in
regulating bodily movements. It is true that the bull's aggressive

charges were stopped for a short period, but there is no
evidence that it was because aggression was inhibited. Rather,
because it was forced to turn in circles every time it came close
to its target, the confused bull eventually stopped charging.
Patients receiving caudate nucleus stimulation also display
various types of stereotyped motor responses. Sometimes all
movement is stopped in an "arrest response," so that a person
instructed to continue tapping a table may be immobilized by
the stimulation with his hand in midair (Van Buren, 1966).
Destruction of the caudate nucleus in cats and other animals has
been reported to produce a syndrome called obstinate progres-
sion, a curious phenomenon characterized by persistent walking
movements even when an animal's head may be wedged into a
corner (Mettler and Mettler, 1942). In humans, movement
disorders such as the spasticity and tremors seen in Parkinson's
disease have frequently been linked to caudate nucleus
pathology. 1

Caudate stimulation has also been reported to cause confu-
sion and to interfere with speech (Van Buren, 1963). There are
several animal studies indicating that caudate stimulation
interferes with the normal habituation of responses to novel
stimuli when they are presented repeatedly, e.g., Deadwyler and
Wyers (1972), and Luria (1973) have suggested that in humans
the caudate nucleus is important for focusing attention because
of its role in selectively inhibiting responses to irrelevant
stimuli. Kirkby and Kimble (1968) reported that rats have
difficulty inhibiting responses in passive avoidance tests follow-
ing damage to the caudate nucleus, and Rosvold, Mishkin, and
Szwarcbart (1958) have concluded that this structure is

1 Plotnik and Delgado (1970) have presented evidence that stimulation of the
caudate nucleus, putamen, gyrus pyriformis, and gyrus rectus may inhibit the threat-
ening grimaces in monkeys that normally followed tail shock. Although only a mini-
mum amount of data were presented, these changes in the monkeys' behavior did not
seem to be accompanied by motor disturbances or general disorientation. Although
the report suggests that stimulation of some structures may inhibit the expression of
aggressive displays at current intensities that do not produce gross motor disturb-
ances, there is no reason to assume that the large number of other functions believed
to be regulated by these brain areas were unaffected.

involved in delayed alternation and visual discrimination perfor-
mance of monkeys.

Many more functions of the caudate nucleus are described in
the scientific literature, but a cataloguing of them all is not
necessary for our present purpose. It should be clear, however,
that we will not advance very far in our attempt to analyze the
contribution of the caudate nucleus to behavior if we restrict
ourselves to listing the complex behaviors affected by electrical
stimulation. What is needed is a testing program designed to
characterize functional changes with increasing precision by
dissecting out the elements common to behaviors appearing to
be very different.

The fact that it is possible to inhibit or evoke different
complex behaviors by electrical stimulation has led some people
to conclude that specific behaviors might be modified by
destroying the neural area around the tip of the stimulating
electrode. Thus, using the electrode implanted in the bull's
caudate nucleus to destroy a portion of this structure would be
expected to alter the aggressive temperament of that animal.
Although the specific experiment has not been done, there is no
reason to believe that this would be the case. Destruction of the
caudate nucleus does not change the aggressive tendencies of
other animals, but it may produce various movement deficits or
impairments on tasks requiring a selective inhibition of sensory
and motor processes and the connections between them. 1
Similarly, if one destroys the hypothalamic area that evokes
aggressive behavior in a cat or rat, under an electrode, no change
in natural aggressivity is induced unless the area destroyed is so
extensive that the animal is incapable of any behavior at all.
Even after surgical isolation of the entire hypothalamus, a cat is
still able to display integrated attack and rage responses when

1 None of this evidence is meant to argue against the possibility that parts of the
caudate nucleus may be more involved in one type of process than in another. It has
been shown that specific parts of the caudate nucleus receive input from the orbital
frontal, the dorsolateral frontal, or the inferotemporal cortex, and the deficits that
follow selective destruction of portions of this complex structure differ accordingly
(Divac, Rosvold, and Szwarcbart, 1967). The behavioral manifestations of these
deficits, however, vary with the demands of the situation.

provoked, as Ellison and Flynn (1968) have demonstrated.
Earlier, Hess described his disappointment at not being able to
modify a behavior elicited by stimulation even after destroying
the tissue around the electrode. He said:

"This step, involving the use of the same electrodes, seemed
to be most promising, inasmuch as we expected that a
comparison of stimulation and destruction effects would
provide us with a reciprocal confirmation in the sense of a plus
or minus effect. In reality, however, the results were disappoint-
ing. Today we know why. Since our procedure aimed for the
greatest possible precision, we often produced only correspond-
ing small foci of coagulation. As is shown by the stimulation
study, however, even the best demarcated 'foci' are relatively
diffuse" (Hess, 1957, p. 43).

Luria (1973, pp. 33-34) commented that localization of
complex functions in specific regions of the brain is always
misleading. What is needed, he said is to "ascertain by careful
analysis which groups of concertedly working zones of the brain
are responsible for the performance of complex mental activity;
what contribution is made by each of these zones to the
complex functional system."

Luria also noted that though it is appropriate to speak of the
secretion of bile as a function of the liver, insulin secretion as a
function of certain cells in the pancreas, and the transduction of
light by photosensitive elements in the retina, when we speak of
such functions as digestion or perception, "it is abundantly
clear that [they] cannot be understood as a function of a
particular tissue." Similarly, Luria (1973) quoted Pavlov on the
question of a "respiratory center." "Whereas at the beginning we
thought that this was something the size of a pinhead in the
medulla . . . now it has proved to be extremely elusive, climbing
up into the brain and down into the spinal cord, and at present
nobody can draw its boundaries at all accurately."

The idea that the brain is organized into discrete compart-
ments whose function corresponds to our social needs is simply
not in accord with reality. The brain does not work that way. A
concept such as aggression is a man-made abstraction and it

therefore should not be expected to exist as a separate entity in
the nervous system. Many parts of the nervous system play roles
in regulating what most of us would label aggressive behavior
and each of these parts also plays a role in regulating other
aggression, and copulation. Even though all of these behaviors
point. These investigators destroyed a small amount of the
hypothalamic tissue in a rat by means of a specially designed
knife and reported changes in eating, drinking, irritability,
aggression, and copulation. Even though all of these behaviors
were not affected equally, the possibility of modifying a large
number of behaviors by destroying even a small amount of
brain tissue is quite clear. In drawing conclusions from brain
stimulation experiments, what is almost invariably overlooked is
that just about every area of the brain is involved in many
different functions and all but the simplest functions have
multiple representation in the brain.

The eagerness to believe that discrete and natural motiva-
tional states such as hunger can be manipulated by brain
stimulation has resulted in a selective perception of even some
of the pioneering work in this field. For example, although Hess
is consistently mentioned as having produced bulimia by
hypothalamic stimulation, it sometimes seems that his classic
papers are not read so often as they are cited. Hess (1957, p.
25) actually said the following: "Stimulation here produces
bulimia. If the animal has previously taken neither milk nor
meat, it now devours or drinks greedily. As a matter of fact,
that animal may even take into its mouth or gnaw on objects
that are unsuitable as food, such as forceps, keys, or sticks. "
(Italics mine.)

It must be recognized that most hungry cats are more
discriminating than Hess's brain-stimulated animals.

In the studies from my own laboratory, it has been shown
that the behavior evoked by brain stimulation is very different
from behavior motivated by natural states. A stimulated animal
may eat one type of food, but not the food it normally eats in
its home cage (fig. 1 ). or it may not eat even the same food if it
is changed in texture, as when food pellets are offered as a

ground mash. Stimulated animals may drink water from a
drinking tube, but not from an open dish (fig. 2), and the taste
preferences of an animal drinking in response to stimulation
differ from those of a thirsty animal (fig. 3). Most important
from the point of view of behavior control (or lack of it), the
elicited behavior may change even in response to identical brain
stimulation. A rat that drinks only in response to stimulation,
for example, may start to eat when stimulated at a later time
(figs. 4, 5). Moreover, the brain sites from which eating and
drinking may be evoked are much more widespread than is
usually implied. There is no anatomically discrete focus for this
phenomenon, although there are brain areas where the proba-
bility of evoking eating and drinking is very low (Cox and
Valenstein, 1969). In 1973 Reis, Doba, and Nathan reached a
similar conclusion. These investigators found that they could
evoke grooming, eating, and predatory behavior (depending on
the intensity of the stimulating current) from almost all
electrodes placed in the fastigial nucleus of the cat's cerebellum.
Since the behaviors invariably appeared in the same order as the
stimulus intensity was increased, regardless of the electrode
placement within the fastigial nucleus, the investigators con-
cluded (op. cit., p. 847): "Thus, it is the intensity of the
stimulus and not the location of the electrode which is one of
the determinants of the identity of the behavior. Second, the
observation that the nature of the behavior evoked from a single
electrode at a fixed stimulus intensity could be changed by
altering the availability of goal objects (such as food or prey) is
another demonstration that the locus of the electrode is not
critical. Thus, our findings suggest that the behavioral responses
from fastigial stimulation are probably not due to excitation of
discretely organized neural pathways."

The conclusion to be drawn from these experiments is
certainly not that stimulation at any brain sites can evoke any
behavior if the contingencies are arranged appropriately or that
stimulation at different sites all evoke the same general state.
These misinterpretations continue to appear in print although
we have made an effort to be clear on these points. Valenstein,

i 5

100%

o 10

0% 0%

MLH

0% 0%

^ Col Dog Food
^ Food Pellets

I Water

Stimulation with

Cot - Dog Food

Removed

ALL GOAL OBJECTS PRESENT

TESTS

CAT -DOG FOOD ABSENT

FIG. 1 . Behavior evoked by brain stimulation in a testing situation involving
choices. During initial 3 tests, rats received brain stimulation in the presence of
commercial cat-dog food, their regular food pellets, and a water bottle. Stimulation
evoked eating cat-dog food only. Then cat-dog food was removed. We assumed that if
stimulation evoked a hunger state animals would readily switch to eating food pellets.
Instead, stimulation gradually evoked drinking with increasing regularity (see fig. 5).
After stimulation evoked regular drinking three additional tests with food pellets and
a water bottle were administered. Animals drank almost every time stimulation was
given. Stimulus parameters were invariably the same. Each test consisted of 20 stimu-
lations (20 sec. duration). Maximum score for any one behavior was 20, but animals
could display more than one behavior during a single 20-second stimulation period.
(Data from Valenstein, Cox, and Kakolewski, 1968b.)

Cox, and Kakolewski (1970, p. 30) said: "We are not suggesting
that any elicited response may substitute for any other, but
rather that the states induced by hypothalamic stimulation are
not sufficiently specified to exclude the possibility of response
substitution." And Valenstein (1969, p. 300) said, "[it] is not
meant to imply that it will not be possible to differentiate the
effects of stimulation at different hypothalamic regions, but
rather that the application of specific terms such as hunger,
thirst and sex may not be justified."

It seems clear that some behaviors are more likely to be

100%

o 10

3
O

90%

M Water in Bottle

I Water in Dish

^^ Food Pellets

Drinking Experience
from Water Dish

Stimulation with

Water Dish and

Food Pellets

WATER DISH ABSENT

WATER BOTTLE ABSENT

TESTS

FIG. 2. Behavior evoked by brain stimulation in a testing situation. During initial
tests, rat drank from water almost every time stimulation was administered, but did
not drink water from a dish or eat food pellets. Afterward, the animal drank all the
water from a dish for three days (this was natural drinking; no brain stimulation was
administered) before periodic stimulation in the presence of the water dish and food
pellets were initiated. It was assumed that if thirst had been induced by stimulation
during initial tests, rat would rapidly switch to drinking water from dish when stimu-
lated. Instead, stimulation gradually evoked eating of food pellets. During three
stimulation tests with water dish and food pellets available, rat did not drink, but ate
food pellets during most of stimulation trials. Stimulus parameters were invariably
the same. Each test consisted of 20 stimulations (20-sec. duration). Maximum score
for any one behavior was 20, but animals could display more than one behavior
during a single 20-second stimulation period. (Data from Valenstein, Kakolewski, and
Cox, 1968.)

interchangeable than others. This probably reflects the role of
the sensory, motor, and visceral changes induced by the
stimulation in channeling behavior in certain directions. Although
these bodily changes do not duplicate natural motivational states,
they do play an important role in determining the types of behav-
ior that will Or will not be seen during stimulation.

100

? 90

K 80 -

o 70 -

60 -

o 50 -

3 40 -

u 30 -

20 -

IVofer

30 % Glucose
N = 9

WATER
DEPRIVED

DRINKING ELICITED BY
HYPOTHALAMIC STIMULATION

FIG. 3. Preference for water and glucose by rats receiving brain stimulation and
the same animals being deprived of water for 48 hours. All rats initially drank water
when stimulated but did not eat. In a two-bottle choice test, they preferred glucose
during brain stimulation, but water when thirsty. (Data from Valenstein, Kakolewski.
and Cox. 1968.)

To date, the direct effects of stimulation have been relatively
neglected. Although it is often stated that stimulation does not
produce behavior changes unless the appropriate stimulus is
available, such changes are actually often neglected even when
the data suggest their importance. For example, the first
description of drinking evoked by brain stimulation contained a
strong suggestion that motor responses may have been more
important in directing behavior than any presumed thirst state.
In his report. Greer (T955. pp. 60-61 ) said:

Stimulation of the animal began 24 hours after the electrodes were
implanted. It was immediately apparent that the animal was under great
compulsion to perform violent "licking" activity when a current was
passed between the hypothalamic electrodes. In response to stimula-
tion, it would stand on its hind legs and run vigorously around the glass
enclosed circular cage, licking wildly at the glass wall. This behavior

would cease immediately upon shutting off the current. If the voltage
were slowly increased, licking would gradually become more vigorous.
With stimulation continuing by timer control, the reaction of the
animal changed during the first night. The water bottle containing 200
ml was found completely empty at 9 a.m. even though it had been
filled at 6 p.m. the previous evening. It was now found that stimulation
would result in violent drinking activity. The non-specific licking
response had been lost. As soon as the current was turned on, the
animal would jump for the water bottle and continue to drink avidly
until the switch was turned off. If the water bottle was removed and
the current then turned on, the rat would go back to its "licking"
behavior of the previous day, but would immediately transfer it to
drinking behavior when the water bottle was replaced.

100% 100% 100%

ALL GOAL OBJECTS PRESENT

WATER REMOVED
TESTS

ALL GOAL OBJECTS PRESENT

FIG. 4. Behavior evoked by brain stimulation in a choice situation. Initially,
animal drank only when stimulation was given (first 3 tests). After periodical
stimulation in the presence of food and a wooden block (for gnawing), but without
water bottle, rat gradually began to eat food pellets. Next three tests demonstrated
that stimulation evoked regular eating. Last 3 tests demonstrated that even when
tested with water bottle present, stimulation elicited eating as well as drinking.
Stimulus parameters were invariably the same. Each test consisted of 20 stimulations
(20-sec. duration). Maximum score for any one behavior was 20, but animals could
display more than one behavior during a single 20-second stimulation period. (Data
from Valenstein, Kakolewski, and Cox, 1968a).

Visceral changes produced by the stimulation may also play a
role in determining the behavior evoked by brain stimulation.
For example, Folkow and Rubinstein (1965) contrasted the
visceral changes produced by hypothalamic stimulation that
evokes eating with those changes produced by electrodes
evoking rage reactions. Among the prominent bodily changes

Coling

Drinking

Stimulation

without

reed

FOOD ond WATER PRESENT

20 25 30 35

ONLY WATER PRESENT
TESTS

45 50

FOOO ond WATER PRESENT

FIG. 5. Gradual development of behavior evoked by brain stimulation. Rat was
tested shortly after first demonstration of eating in response to stimulation. In more
than 10 successive tests, eating was evoked by brain stimulation with increasing
regularity. Although water was available, stimulation never evoked drinking. Animal
was then given periodic stimulation for one-week period until it started to drink in
response to stimulation. In more than 40 tests, rat drank in response to stimulation
with increasing regularity. During last 10 tests, both food and water were present.
During most tests rat ate and drank when stimulation was administered, although
drinking gradually became dominant evoked response. Stimulus parameters were invar-
iably the same. Each test consisted of 20 stimulations (20-sec. duration). Maximum
score for any one behavior was 20, but animals could display more than one behavior
during a single 20-second stimulation period. (Data from Valenstein, 1971.)

produced by stimulation that caused rats to eat were a marked
increase of intestinal motility and change in stomach volume
plus mild increases in blood pressure and heart rate. The pattern
was different when rage was evoked; intestinal and gastric
motility were inhibited, and the blood pressure and blood
distribution patterns differed from those produced by elec-
trodes that evoked eating.

Ball (1974) also stressed the importance of visceral changes
for evoked eating. In rats displaying this response, Ball
sectioned the vagus nerve at a point close to where it innervates
the stomach. He reported that the stimulus threshold for
elicitation of eating was raised significantly even after the
animals recovered from surgery and were eating the normal
amount of food in their home cage. Even though the thresholds
increased, it was clear that the visceral changes controlled by

this branch of the vagus nerve were not necessary for the
evoked behavior, as stimulation continued to evoke eating.
Similarly, as noted earlier, Reis, Doba, and Nathan (1973)
reported that electrical stimulation of the rostral fastigial
nucleus of the cat's cerebellum elicited either grooming,
feeding, or killing of a rat, depending on the intensity of
stimulation used. The magnitude of the cardiovascular responses
(heart rate and blood pressure) differed for each of the three
behaviors evoked, but the behaviors were still displayed after
these visceral responses were blocked by an injection of
phentolamine. It is evident from the many important studies by
Flynn and his colleagues (see Flynn, Edwards, and Bandler,
1971) that brain stimulation produces many different sensory,
motor, and visceral changes. Apparently the blocking of one or
two of these changes is not likely to be very disruptive once an
elicited behavior has become established. These bodily changes,
however, may play an important role in channeling behavior
during the initial brain stimulation experience.

In addition to producing bodily changes, the positive or
aversive motivational effects evoked by brain stimulation may
also serve to channel behavior and determine which behaviors
are interchangeable. Plotnik (1974) summarized the motiva-
tional consequences of 174 brain stimulation sites in monkeys.
The motivational effects were determined by tests that mea-
sured whether an animal sought out escaped from or was
indifferent to the brain stimulation. It was found that 1 17 sites
were neutral, 22 were positive or rewarding, and 35 were
aversive or negative. All 14 points that elicited aggressive
behavior directed at other monkeys had aversive motivational
properties, although the converse was not true. Plotnik views
the elicited aggression as "secondary aggression" produced by
reaction to an aversive stimulus. In such cases, it would be as
misleading to conclude that there was a direct relationship
between natural aggression and the brain site stimulated as there
would be to conclude the same about the soles of the feet
because an electric shock delivered to them produces fighting
between animals caged together. The point is well illustrated by

Black and Vandcrwolf (1969, p. 448), who reported that a
foot-thumping response could be evoked in the rabbit by
stimulation of diverse brain sites (in the hypothalamus, thala-
mus, central gray, septum, reticular formation and fornix-
fimbria). Rather than postulate the existence of a complex
"thumping circuit" in the rabbit brain, they noted that
thumping could be elicited by foot shock and concluded that
"thumping behavior in the rabbit is a fear or pain response. "

The significance of the motivational properties of brain
stimulation is made clearer by distinguishing predatory from
aggressive behaviors. In cats and rats, hypothalamic stimulation
has evoked both types of behaviors. In these animals, a
predatory, stalking behavior (called "quiet biting attack" in the
rat), which is well directed at an appropriate prey, has been
distinguished from a diffusely directed "affective rage attack"
(Wasman and Flynn, 1962; Panksepp, 1971). Stimulation at
sites that evoke the predatory (or appetitive) behavior has been
shown to also evoke positive or rewarding effects (Panksepp,
1971), whereas stimulation at sites evoking "affective attack"
has been demonstrated to be aversive (Adams and Flynn, 1966).
In primates, the elicited aggression is intraspecific, resembling
fighting rather than predatory behavior, and is evoked pri-
marily, if not exclusively, by stimulation having aversive
motivational properties. 1 Although the evidence is inadequate,
aggression provoked by brain stimulation in humans also seems
to occur only in cases of stimulation having aversive conse-
quences (see Valenstein, 1973, for a review of this literature).
Considerations such as these, suggesting that certain behaviors
are compatible with aversive and others with positive states,
may set limits on the behaviors that can be evoked by a
particular brain electrode.

Although the somatic and motivational effects produced by

1 Robinson, Alexander, and Browne (1969) reported one instance where stimula-
tion elicited aggressive attacks on another monkey and also supported self-stimula-
tion behavior. This suggests that brain stimulation that elicits intraspecies aggression
may be motivationally positive. However, as self-stimulation was tested with brief
(0.5 sec.) stimulus trains and aggression was elicited by relatively long (10-40 sec.)
stimulus trains, this exception may be more apparent than real.

brain stimulation make it more likely that one group of
behaviors will be evoked rather than another, these factors are
by no means sufficient to determine completely the specific
behavior displayed or the motivational states induced. Environ-
mental factors and individual or species characteristics can also
be very important determinants. An experiment from my own
laboratory demonstrates this point and also illustrates how
easily one can be misled by first impressions in brain stimula-
tion experiments.

Figure 6 illustrates a two-compartment chamber used to test
the behavior of rats receiving rewarding hypothalamic stimula-
tion (Phillips et al., 1969). The equipment was so arranged that
when the rat interrupted the photo cells on the right side of the
chamber, brain stimulation is turned on and remained on until
the animal interrupted the photo cell in the left compartment.
In an amazingly short time, the rat learned to play the game,
running to the right side and turning the stimulation on for a
period, then running to the opposite side and turning the
stimulation off. This behavior was repeated rapidly over and
over again. The rat was stimulating its own brain and apparently
enjoying it— at least it continued doing it.

At this point, we placed food pellets on the right, the
stimulation side. After a brief period, the rat started to pick up
the pellets when stimulated and carry them (as pictured in fig.
6) to the opposite side of the chamber, where they were
dropped as soon as the brain stimulation was turned off. We
were fascinated by this unexpected turn of events, as it seemed
possible that we had stumbled on a region of the rat's brain that
regulated food-hoarding behavior. At least that is what we were
thinking until we investigated a little further. When we
substituted rubber erasers and pieces of dowel sticks, the rat
carried them just as readily. If we mixed the edible and inedible
objects together, the rat did not discriminate between them. It
carried both. This was a very strange type of food-hoarding
behavior! Next we placed some rat pups on the right side and
found that these also were carried to the other side. It dawned
on us that had we started with the rat pups and gone no farther,

we would have been convinced thai we were activating the brain
structures that controlled the pup-retrieval component of
maternal behavior. Probably we would have found it difficult to
resist speculating about the significance of the fact that males
carried pups as readily as females.

FIG. 6. Rat carrying wooden dowel stock from stimulation (right) to nonstimula-
tion (left) side of test chamber. (In this variation of the basic experiment, animals
were given a choice between receiving hypothalamic stimulation with or without the
opportunity to carry objects; they chose the former.) (Data from Phillips et al., 1969:
figure reproduced from Valenstein, Cox, and Kakolewski, 1970.)

Once the rats started carrying objects regularly, they would
pick up and carry almost anything in response to the stimula-
tion. When stimulated, the compulsion to carry became so
strong that the rats carried parts of their own bodies when all
the objects were removed. A rat picked up its tail or a front leg
with its mouth and carried it over to the other side where it was
"deposited" as soon as the stimulation was turned off.

Finally, we found that if the very same stimulation was
delivered to the rat's brain under different conditions, objects
were no longer carried. We programmed the equipment to
deliver the same temporal pattern of stimulation the rat had
previously self-administered, controlled now by a clock rather
than by the rat's position. This procedural change resulted in
the possibility that the animal could be stimulated any place in
the test chamber rather than the stimulation being turned on

and off consistently in different parts of the chamber. Under
these conditions the identical electric stimulus, delivered to the
same brain site through the same electrode, no longer evoked
object-carrying even if the animal was directly over several of
the objects when the stimulation was turned on.

We believe that the answer to this puzzling phenomenon lies
partly in the rat's tendency to carry objects (food, pups, even
shiny objects) from an open field, where the rat is vulnerable
and therefore highly aroused, back to the relatively secure and
calming environment of the nest site. When stimulation is
delivered regularly in certain parts of the rat's life space and
turned off regularly in other parts, it not only produces
alternating arousal and calming states, but links these states to
specific parts of the environment. In addition, because rats
prefer to turn off even rewarding brain stimuli after a period of
time (Valenstein and Valenstein, 1964), they are forced to
move back and forth in the test chamber. Taken all together, we
may have inadvertently duplicated all the internal and external
conditions that exist when a rat makes repeated forays from its
nest site to the outside world.

Admittedly, this explanation is speculative. It is clear,
however, that the behavior produced by stimulation is not
determined in any simple fashion by the location of the
electrode in the brain. (Actually, we achieved the same results
with electrodes in different rewarding sites.) The behavior
produced by the stimulation can only be understood by
considering the natural propensities of the rat in the environ-
mental conditions in which it is tested.

BRAIN STIMULATION IN HUMANS AND OTHER PRIMATES

If the response to brain stimulation is variable in inbred rats,
it is certainly much more variable in monkeys and humans. In
monkeys, for example, brain stimulation may initiate drinking
when the animal is confined to a restraining chair. However,
when the stimulation is administered when the monkeys are in a
cage and not restrained, they do not drink, even though they

may be sitting within inches of the water dispenser when the
stimulation is administered (Bowden, Galkin, and Rosvold, In
press). In humans, brain stimulation may evoke general emo-
tional states that are somewhat predictable in the sense that
certain areas tend to produce unpleasant feelings and other
areas tend to produce positive emotional states. Patients may
report feeling tension, agitation, anxiety, fear, or anger, or they
may describe their feelings as being very pleasant or relaxed.
Different patients report different feelings from stimulation of
what is presumed to be the same brain area, and the same
person may have very different experiences from identical
stimulation administered at different times (see Valenstein,
1973, for a review of this literature). The impression that brain
stimulation can evoke the identical emotional state repeatedly
in humans is simply a myth, perhaps perpetuated in part
because of its dramatic impact. Janice Stevens et al. (1969, p.
164) stressed this variability: "Subjective changes were elic-
itable in similar but not identical form repeatedly on the same
day, but often were altered when stimulation was carried out at
the same point on different days. "

Many people have the impression that the results of brain
stimulation are predictable because of the reports that the same
visual hallucinations and memories can be evoked repeatedly by
brain stimulation. It is true that Wilder Penfield, who operated
on the temporal lobes of patients suffering from intractable
epilepsy, had emphasized that electrical stimulation of this
brain region may repeatedly evoke the same memory. Consider-
able excitement was generated by reports that these evoked
memories had the fidelity of tape recording playbacks of past,
forgotten experiences. Indeed, on the basis of these reports, a
few psychoanalysts began to speculate about the neural basis of
repressed memories (Kubie, 1953). What was generally over-
looked, however, was that Penfield had reported that the same
response could be evoked within a minute or two, but a
different response was obtained after a longer period (see
Penfield and Perot, 1963). The similarity of this conclusion to
that of Stevens et al. (1969) is apparent. Moreover, recent

studies have made it clear that the occurrence of these evoked
memories is rare and when they do occur it can usually be
shown that they were determined by what was on the patient's
mind or some other aspect of the situation when stimulation
was administered (Van Buren, 1961; Mahl et al., 1 964).

Even relatively simple motor and sensory responses to
stimulation of specific areas of the cerebral cortex of primates
may vary with time and individuals. When Leyton and
Sherrington (1917) reported their observations following corti-
cal stimulation of the chimpanzee, orang-utan, and gorilla, they
noted considerable evidence of "functional instability of corti-
cal motor points." Not only did thresholds vary and stimulation
of a particular brain site produce either extension or at other
times flexion of the same joint, but the muscles involved
sometimes also changed. Leyton and Sherrington reported that
often a particular response became dominant and was elicited
from a variety of cortical points that had previously given very
different responses. They also observed that stimulation of the
same cortical points produced different responses from differ-
ent individuals and even from opposite hemispheres within the
same individual. This is not to deny that there was general
agreement as to the parts of the frontal cortex most likely to
produce movement of some kind in specific muscle groups, but
Leyton and Sherrington emphasized that the details of the
movements would not be the same if the experiment were
repeated. Observations of this type have also been made
following stimulation of the human cortex. Penfield and
Boldrey (1937, p. 402) noted that stimulation at a point on the
post-central gyrus, which does not elicit a particular response,
may gain this capability if it is tested after stimulating a brain
point that does evoke the response. Similar observations of
variation of responses have been reported following electrical
stimulation of sensory cortical areas in humans. Penfield and
Welch (1949), for example, noted that if a brain site evoked
sensations seeming to originate in the thumb, the same
stimulation might later evoke sensations experienced as coming
from the lips if the stimulation had been preceded by activation

of another site that evoked lip sensations. These authors have
called such variability "deviation of sensory response." Libet
(1973) discussed the variability in human response to electrical
stimulation in more detail.

It is totally unrealistic to believe that stimulation of a
discrete point in the brain will invariably elicit the same
memory, emotional state, or behavior. The changes produced
by the stimulation depend upon what is going on in the rest of
the brain and in the environment at the time. The understand-
able need in science to eliminate variability and demonstrate
control over phenomena may, when applied to the study of the
brain, distort reality by concealing the very plasticity that is an
essential aspect of adaptive behavior.

CONTROL OF HUMAN BEHAVIOR: FACT AND FANTASY

No discussion of electrical brain stimulation and behavior
control would be complete without considering the existence of
rewarding brain stimulation. As everyone surely knows by now,
Olds and Milner (1954) accidentally discovered about 20 years
ago that electrical stimulation of certain brain structures can
serve as an effective reward for rats. Subsequent studies of the
behavior of rats and other animals indicated, in many different
ways, that pleasurable sensations can be evoked by brain
stimulation (see Olds, 1973). No other single discovery in the
brain-behavior field has produced more theoretical speculation
than the phenomenon that animals are highly motivated to
stimulate their own brains. Clarke's reaction (1964, pp. 200-
201) to this discovery is representative:

Perhaps the most sensational results of this experimentation, which
may be fraught with more social consequences than the early work of
the nuclear physicists, is the discovery of the so-called pleasure or
rewarding centers in the brain. Animals with electrodes implanted in
these areas quickly learn to operate the switch controlling the
immensely enjoyable electrical stimulus, and develop such an addiction
that nothing else interests them. Monkeys have been known to press the
reward button three times a second for eighteen hours on end,
completely undistracted either by food or sex. There are also pain and

punishment areas of the brain; an animal will work with equal
singlemindedness to switch off any current fed into these.

The possibilities here, for good and evil, are so obvious that there is no
point in exaggerating or discounting them. Electronic possession of
human robots controlled from a central broadcasting station is
something that even George Orwell never thought of, but it may be
technically possible long before 1984.

In part, because the pleasurable reactions have been produced
by direct stimulation of the brain and involve electronic
gadgetry, there is a tendency to conjure up images of "pure
pleasure 1 ' that are completely irresistible. It should surprise no
one that science fiction writers have seized this phenomenon as
a theme for their stories. In Larry Niven's story (1970), for
example, the presumed omnipotence of rewarding brain stimu-
lation is at the very center of the "perfect crime." The story
takes place in the year 2123 and Owen Jennison's body has just
been discovered under conditions that appear to indicate a
suicide, but the death actually was the result of a carefully
planned murder:

Owen Jennison sat grinning in a water stained silk dressing gown. A
month's growth of untended beard covered half his face. A small black
cylinder protruded from the top of his head. An electric cord trailed
from the top of the cylinder and ran to a small wall socket.

The cylinder was a droud, a current addict's transformer.

It was a standard surgical job. Owen could have had it done anywhere.
A hole in his scalp, invisible under the hair, nearly impossible to find
even if you knew what you were looking for. Even your best friends
wouldn't know, unless they caught you and the droud plugged in. But
the tiny hole marked a bigger plug set in the bone of the skull. I
touched the ecstasy plug with my imaginary fingertips, then ran them
down the hair-fine wire going deep into Owen's brain, down into the
pleasure center.

He had starved to death sitting in that chair.

Consider the details of the hypothetical murder. Owen Jennison is
drugged no doubt-an ecstasy plug is attached — He is tied up and
allowed to waken. The killer then plugs Mr. Jennison into a wall. A
current trickles through his brain, and Owen Jennison knows pure
pleasure for the first time in his life.

He is left tied up for, let us say, three hours. In the first few minutes he
would be a hopeless addict.

No more than three hours by our hypothesis. They would cut the ropes
and leave Owen Jennison to starve to death. In the space of a month
the evidence of his drugging would vanish, as would any abrasions left
by ropes, lumps on his head, mercy needle punctures, and the like. A
carefully detailed, well thought out plan, don't you agree?

The readiness to believe that artificial stimulation of the
brain can evoke such intense and irresistible pleasures reveals
more about our desires than about our brain. Routtenberg and
Lindy (1965) did demonstrate that some rats actually starved
themselves to death because they continued to stimulate their
brains rather than eat. However, one can be terribly misled by
the popular accounts of this experiment. In the actual experi-
ment, rats with electrodes implanted in rewarding brain
structures were given only one hour a day to press a lever for
food. It was necessary for them to eat during that hour in order
to stay alive. After the rats were on this feeding schedule for a
period, they were given a second lever that offered brain
stimulation as a reward. Some of them spent so much time on
this second lever that they did not receive sufficient food to
keep them alive until the next day's hourly session. This is quite
different from the picture most people have in mind about what
took place. In the special conditions of a brief test designed to
emphasize the controlling power of brain stimulation, some of
the rats were apparently not able to anticipate the consequences
of choosing the brain stimulation lever. Under conditions
providing rats with free access to brain stimulation and food,
they never starve themselves. In fact, they eat their usual
amount of food (Valenstein and Beer, 1964).

Rewarding brain stimulation is not equally compelling for all
species. In humans, it does not seem capable of inducing an
irresistible, pleasurable experience. Robert Heath, who is
probably more experienced than anyone else with the pleasura-
ble reactions brain stimulation can evoke in humans, has
commented that it does not seem able to induce a euphoria
equal to that produced by drugs (personal commun.). This is

not to deny that patients have reported feeling considerable
pleasure during brain stimulation or that they were willing to
repeat the experience, particularly after receiving the impression
that this was part of the therapeutic program. (See Valenstein,
1973, for a review of the reports of pleasure evoked by brain
stimulation in humans.) Brain stimulation has evoked orgasms,
but there is a tendency to attach too much significance to this.
It is usually overlooked that, as with masturbation, brain
stimulation that produces an orgasm does not continue to be as
pleasurable afterward.

The emotional state induced in humans by brain stimulation
varies with the emotional and physical condition of the patient.
(Heath, John, and Fontana, 1968, p. 168) stated that "When the
same stimulus was repeated in the same patient, responses
varied. The most intense pleasurable responses occurred in
patients stimulated while they were suffering from intense pain,
whether emotional and reflected by despair, anguish, intense
fear or rage, or physical, such as that caused by carcinoma. The
feelings induced by stimulation of pleasure sites obliterated
these patients' awareness of physical pain. Patients who felt well
at the time of stimulation, on the other hand, experienced only
slight pleasure. " (Italics mine.)

The existence of circuits in the brain that can induce both
pleasure and arousal may be telling us something important
about neural mechanisms that have evolved to help focus
attention, to increase involvement in a task, and to facilitate the
consolidation of memories (see discussion in Valenstein, 1973,
pp. 40-44). There are speculations that malfunctioning of these
reward circuits is responsible for such psychiatric conditions as
depression and schizophrenia (see Stein, 1971). Such specula-
tion leads to more research that ultimately will increase our
understanding of how the brain regulates behavior. This is very
unlikely to be the consequence of the proposals to use brain
stimulation to control behavior.

It would be difficult to fabricate a better example of the
distortions that can result from a preoccupation with behavior
control than that contained in a proposal, apparently seriously

advanced, by [ngraham and Smith ( 1972). These two criminolo-
gists suggested that techniques are available for maintaining a
surveillance on paroled prisoners and for controlling their
behavior. They propose that implanted devices could be used to
keep track of the location of the parolee and his physiological
state while remotely operated brain stimulation could deliver
either rewards or punishments or it could control behavior in
other ways. For example, Ingraham and Smith (1972, p. 42)
suggested the following scenario.

"A parolee with a past record of burglaries is tracked to a
downtown shopping district (in fact, is exactly placed in a store
known to be locked up for the night) and the physiological data
reveals an increased respiration rate, a tension in the muscula-
ture and an increased flow of adrenalin. It would be a safe
guess, certainly, that he was up to no good. The computer in
this case, weighing the probabilities, would come to a decision
and alert the police or parole officer so that they could hasten
to the scene; or, if the subject were equipped with an implanted
radiotelemeter, it could transmit an electrical signal which could
block further action by the subject by causing him to forget or
abandon his project."

It is impossible to be certain, but it seems unlikely that
anyone would approve such a plan. The more serious problem is
the amount of creative energy diverted from the search for
realistic solutions to important social problems by this type of
thinking. It sometimes seems that difficulties in implementing
necessary social changes encourage people to search for solu-
tions in a fantasy world.

Hopefully, it is clear by now that the responses that can be
evoked from stimulating discrete brain areas are too variable
and affect too many different functions to be useful in
behavior-control schemes. The evoked behavior depends on
what is going on elsewhere in the brain, individual and species
characteristics, and is very much influenced by situational
factors. Those who prefer to think only in terms of control may
be very disappointed to learn this. Those who think that the
basic concern of science is understanding may find it useful to

be reminded of the complex relationship between brain and
behavior.

It is not surprising that biological solutions to social problems
have been discussed most frequently in the context of control-
ling violence. This discussion, and some actual proposals, have
taken very different forms. In his address to the American
Psychological Association mentioned earlier, Kenneth Clark also
stated:

Given the urgency of the immediate survival problem, the psychological
and social sciences must enable us to control the animalistic, barbaric
and primitive propensities in man and subordinate these negatives to
the uniquely human moral and ethical characteristics of love, kindness
and empathy. We can no longer afford to rely solely on the traditional
prescientific attempts to contain human cruelty and destructiveness.

Given these contemporary facts, it would seem that a requirement
imposed on all power-controlling leaders, and those who aspire to such
leadership, would be that they accept and use the earliest perfected
form of psychotechnological, biochemical intervention which would
assure their positive use of power and reduce or block the possibility of
using power destructively. It would assure that there would be no
absurd or barbaric use of power. It would provide the masses of human
beings with the security that their leaders would not sacrifice them on
the altars of their personal ego (Presidential Address, American
Psychological Association, 1971).

Undoubtedly Kenneth Clark is seriously concerned about
possible misuse of the enormous capabilities for destruction
that exist. His speech and the types of solutions he proposes
make it apparent that he has been greatly influenced by the
experiments interpreted as revealing discrete neural circuits
regulating aggression. Stripped to its essentials, his proposal
appears as a modern variant of phrenology, a belief that the
brain is organized into convenient functional systems that
conform to our value-laden categories of behavior. Clark seems
to believe that we need only to exorcise those critical regions of
the brain that are responsible for undesirable behavior, or to
suppress them biochemically, and goodness will dominate.
Mankind will be saved by a "goodness pill." The great impact of
the many distorted descriptions of the power of brain control

techniques becomes especially evident when even social sci-
entists accept the questionable hypotheses that wars are mainly
caused by man's animal-like aggressive tendencies and that
biological intervention offers a practical way to prevent them.
Clark has not suggested any specific biological intervention, so
it is not possible to discuss his proposal in any detail. The
situation is different with the proposal advanced by Vernon
Mark and Frank Ervin.

Mark and Ervin (1970) stressed the magnitude of the
problem of violence in the United States and the belief that a
biological approach can make a significant contribution toward
finding a solution. The following are typical of a number of
statements from their book.

"Violence is, without question, both prominent and preva-
lent in American life. In 1968 more Americans were the victims
of murder and aggravated assault in the United States than were
killed and wounded in seven-and-one-half years of the Vietnam
War; and altogether almost half a million of us were the victims
of homicide, rape, and assault."

They introduced their book (1970) with the Preface that
"We have written this book to stimulate a new and biologically
oriented approach to the problem of human violence." In the
foreword to the book, William Sweet, a neurosurgeon affiliated
with Harvard University and the Massachusetts General Hospital
and a frequent collaborator of Mark and Ervin, expressed "the
hope that knowledge gained about emotional brain function in
violent persons with brain disease can be applied to combat the
violence-triggering mechanisms in the brains of the non-
diseased." Clearly, a biological solution to the problem of
violence is sought.

Mark and Ervin suggested that abnormal brain foci in the
amygdala are responsible for a significant amount of violent
crimes. They believe that these abnormal foci often respond to
internal and external stimuli by triggering violent behavior.
Mark and Ervin have implanted stimulating electrodes in
patients that display a history of episodic violence and claim to
be able to locate the "brain triggers" by determining the area

from which violent behavior can be evoked. The treatment
consists of destroying the area believed to be responsible for the
abnormal behavior.

The relevance of temporal lobe structures for aggressive
behavior can be traced back to the seminal studies of Kliiver
and Bucy (1939), although there were several earlier reports
that contained similar observations (for example, Brown and
Shafer, 1888; Goltz, 1892). Most investigators now believe that
the temporal lobes and particularly the amygdala nuclei play an
important, although complex, role in the expression of aggres-
sion, but Kliiver and Bucy and all subsequent investigators have
emphasized the very many different behavioral changes that
follow destruction of this brain region in animals (see Valen-
stein, 1973, pp. 131-143).

In addition to a "taming" of monkeys (and other animals)
after temporal lobe ablation, hypersexuality, increased orality,
and a so-called psychic blindness 1 have also been observed.
Others have emphasized the emotional "flatness" of the
amygdalectomized animal (e.g., Schwartzbaum, 1960). The
behavior changes may take very different forms, even diametri-
cally opposite expression, under different circumstances. Amyg-
dalectomized monkeys may become less aggressive toward man,
but as Rosvold, Mirsky, and Pribram (1954) reported, the
changes in dominance patterns between animals may be more
dependent on the history of their social interactions than on the
particular brain area destroyed.

Arthur Kling and his colleagues have recently reported even
more striking evidence of the fallacy of describing complex
change in response tendencies by such shorthand expressions as
"increased tameness" (Kling, Lancaster, and Benitone, 1970;
Kling, 1972). Kling captured and amygdalectomized wild
monkeys in Africa and on Caijo Santiago near Puerto Rico.
Control monkeys that were captured and released rejoined their

1 "Psychic blindness" refers to a loss of higher integrative visual functions rather
than to a loss in visual acuity.

troupe although sonic initial fighting was necessary. Before
they were released, the amygdalectomized monkeys seemed
tamer when approached by the experimenters, but when
released into their own troupe they were completely unable to
cope with the complexities of monkey social life. The behavior
of the amygdalectomized monkeys was often inappropriate.
Sometimes they displayed aggression toward dominant animals,
a trait never exhibited before. In not too long a period, all the
amygdalectomized monkeys either were driven from or re-
treated out of the troupe and eventually either died of
starvation or were killed by predators. These observations
demonstrate the multiplicity of behavioral changes that usually
occur following brain lesions and the dependency of these on
environmental conditions. In this context, it is interesting that
the compulsive sexual mounting commonly observed in amyg-
dalectomized monkeys housed in the laboratory was not seen
under natural conditions.

The results of amygdalectomy in humans have been less
systematically studied. These operations have been performed
on patients exhibiting aggressive, hyperkinetic, and destructive
behavior, usually (but not always) accompanied by temporal
lobe epilepsy. While hypersexuality and orality have been
observed to occur postoperatively in humans, most neurosur-
geons claim these symptoms are rare and when they occur they
subside after several months (see Valenstein, 1973, pp. 209-233,
for a review of the clinical literature). Although "psychic
blindness" has not been reported, there exist only a few serious
studies of intellectual changes following amygdalectomy in
humans. In one study, Ruth Andersen (1972) tested 15 patients
after amygdalectomy, and even though 13 of them had
undergone only unilateral operations, she reported evidence of a
loss of ability to shift attention and respond emotionally.
Anderson (1972, p. 182) concluded, "Typically the patient
tends to become more inert, and shows less zest and intensity of
emotions. His spontaneous activity tends to be reduced and he
becomes less capable of creative productivity.

"With these changes in initiative and control of behavior, our

patients resemble those with frontal lesions. It must be pointed
out, however, that the changes are very discrete and there is no
evidence of serious disturbance in the establishment and
execution of their major plans of action.

"Presumably he will [function best] in well-structured
situations of a somewhat monotonous and simple character."

Typically, amygdalectomy in humans involves destruction of
an appreciable proportion of this structure. For example,
Heimburger, Whitlock, and Kalsbeck (1966) and Balasubra-
maniam, Kanaka, and Ramamurthi (1970) estimated that they
had destroyed more than 50 percent of the amygdala on each
side. In view of the animal literature and Ruth Andersen's
observations, one might suspect that had adequate postopera-
tive testing been generally used, intellectual and emotional
deficits would have been detected more often. Mark and Ervin
(1970, p. 70) implied that their lesions need not be large
because of the use of stimulating electrodes to locate the
discrete focus that is triggering the violence. They argued that
postoperative deficits would be minimized by the smaller, more
selective stereotaxic lesions their technique makes possible. For
example: "tiny electrodes are implanted in the brain and used
to destroy a very small number of cells in a precisely
determined area. As a surgical technique, it has three great
advantages over lobectomy: it requires much less of an opening
in the surfaces of the brain than lobectomy does; it destroys less
than one-tenth as much brain tissue; and once the electrodes
have been inserted in the brain, they can be left without harm
to the patient until the surgeon is sure which brain cells are
firing abnormally and causing the symptoms of seizures and
violence."

It is important, therefore, to examine critically the validity of
the claim that electrical stimulation is a reliable means of
locating a "brain trigger of violence."

A few years ago, while studying the elicitation of behavior by
hypothalamic electrodes, we noticed an interesting trend (Cox
and Valenstein, 1969). In each of the rats we had implanted
two electrodes, one on each side of the midline, but usually not

symmetrically placed. We observed that in a number of animals
the same response was evoked from very different placements,
whereas in other animals either different or no specific behavior
was elicited from electrodes that often seemed to be in the same
locations (fig. 7). We concluded that within certain anatomical
limits, a "prepotent response" tendency of the animal (Valen-
stein, 1969) appeared to be a more important determinant of
the behavior evoked than the exact location of the electrode in
the brain.

"'"^0^ 44AAL

15AAL 1SAAR

Behavior: BotiaalE).

Drink ;-.j ■; D), Gnawing (G)

Left(L) Electrode

Electron

Strong D

Strong

Intermediate E+D

iit< E*D

Intenediite E

! i -it* E

Strong D

Intermediate G

Strong G

Strong D

FIG. 7. Illustration of different anatomical locations for two electrodes that
evoked the same behavior in a given animal. (Data from Valenstein, Cox, and Kako-
lewski, 1970. Brain diagrams from Konig and Klippel, 1963.)

Many were skeptical of our conclusion and cited examples
from the literature or from their own laboratory experiences
that demonstrated that two electrodes could evoke different
behaviors in the same animal. We had never denied this, but had
argued that many electrodes evoke states that are sufficiently

similar, yet not specifically identifiable, so that the stimulated
animal's behavioral characteristics become a major determinant
of the effects produced by stimulation. Additional information
has been accumulating supporting our impression. In a recent
study using monkeys, it was noted that drinking was elicited
initially in some by only a few electrodes, but over time an
increasing number of electrodes situated at different brain sites
gained the capacity to evoke drinking. Stimulation at an equally
varied distribution of sites in other monkeys did not evoke
drinking. Some monkeys seem to respond to brain stimulation
at many different sites by drinking, whereas others do not
(Bowden, Galkin, Rosvold, In press). A similar conclusion may
be drawn from an earlier study by Wise (1971) in which rats
were implanted with electrodes capable of being moved up and
down within the brain (fig. 8). It was found that in some rats
eating and drinking were continuously evoked as the electrode
was advanced over a large dorsoventral portion of the hypo-
thalamus, but in other rats, these behaviors were not observed
in response to stimulation at any site (fig. 9).

FIG. 8. Sketch of an electrode assembly that can be raised and lowered in the
animal's brain. (See Wise, 1971 , for details.)

Panksepp (1971, p. 327) has also provided information that
supports our "prepotency hypothesis. " He has studied the elicita-
tion of mouse-killing responses in rats and has concluded thai

FIG. 9. Path of electrodes used to explore brain for regions evoking eating and
drinking. Electrodes were advanced in 0.5 mm. steps descending along the path of the
tract. Upper sections show paths of electrode penetrations that did not evoke eating
or drinking. In lower sections, both eating and drinking were evoked from all posi-
tions between upper and lower circles. Each electrode was placed in a different
animal. (See Wise, 1971, for more details.)

the ability to elicit mouse-killing by stimulating the brain of a
rat ". . . interacted with the behavioral typology of individual
animals, animals normally inclined to kill mice were more likely
to kill during hypothalamic stimulation than nonkillers. Thus,
the electrically elicited response was probably not determined by
specific functions of the tissue under the electrode but by the
personality of the rat."

In regard to humans, Kim and Umbach (1973) reported the
effects of stimulating the amygdala of aggressive and nonaggres-
sive patients. They concluded that during amygdala stimulation
of aggressive patients ''aggressiveness increased, whereas no
aggressive reaction was observed in non-violent cases. Thus the
amygdaloid complex seems not to be specific for anxiety alone
or for aggression alone, and shows no specificity of the
subnuclei for these emotional states."

There is little reason, therefore, to believe that brain
stimulation is a reliable technique for locating discrete foci that
trigger violence even if such foci exist. In the violence-prone
patients sent to Mark and Ervin, violence can be triggered by a
great number of brain stimulation sites and probably also by a
pinch on the skin. The ability of stimulation techniques to
ferret out a "critical focus" is far from what it has been touted
to be. Indeed, the fact that Mark and Ervin found it necessary
to make bilateral lesions to produce any significant effect
strongly suggests that no "critical focus" was found. Also
supporting this interpretation is the fact that the bilateral
lesions are usually made progressively larger until the desired
behavior change is believed to have been achieved. Although
Mark and Ervin have presented their approach very seductively
by implying that they can locate and eliminate small and
discrete "brain triggers of violence," in actual practice they
seem to be performing "standard" bilateral amygdalectomies.

There is little doubt that there are well-documented cases
where the onset of assaultive behavior can be traced to temporal
lobe damage. There is also little doubt that there are cases
where, by all reasonable standards, surgery has led to consider-
able improvement in behavior (Gloor, 1967). There has,
however, been a gross exaggeration of the amount of violence
that can be attributed to brain pathology. The evidence
presented by Mark and Ervin is extremely weak. It consists
mainly of a recitation of parallel statistics on the numbers of
murders, rapes, assaultive acts, automobile accidents, and
assassinations, on one hand, and the number of cases of
epilepsy, cerebral palsy, mental retardation, and other indica-

tions o\ brain damage, on the other. Not only arc no causal
connections established, but the statistical evidence does not
support the conclusion that the correlation of brain damage and
violence is high. 1 Mark and Ervin have also bolstered their
general argument by implying that brain pathology was the
cause in such dramatic and violent incidents as the Charles
Whitman shooting from the University of Texas tower. 2

Totally neglected in their description was Whitman's personal
history, which could readily have provided an explanation for
his violence without any brain pathology. Nor was there any
mention that Whitman's carefully laid plans did not conform to
the pattern of sudden, unprovoked, episodic violence that Mark
and Ervin have described as characteristic of those with
abnormal brain foci. It may be relevant to point out that
according to the newspapers. Whitman's brother was shot to
death in a barroom dispute not too long ago. Is it likely that a
temporal lobe tumor was the cause here, too?

There is a danger that the frustration produced by the
inability to effectively reverse the accelerating rate of violence
will cause those whose minds run toward simplified behavior-
control schemes to accept the delusion that biological solutions
are available for what are primarily social problems. The varying
amount of violence prevalent at different times and in different
societies makes it clear that violence is primarily a social

1 The older neurological and psychiatric literature often contained statements that
epileptics, particularly temporal lobe epileptics, are prone to violence. Most neurolo-
gists today refute the earlier figures. Current estimates of the incidence of violence
among epileptics ranges between 1 and 4 percent and if corrections are made for age
(onset of temporal lobe epilepsy is later than for other epilepsies) the relationship is
no higher for the temporal lobe subgroup. Rodin (1973) induced seizure in 150
epileptic patients using the EEG activating drug, bemigride. He reported that there
was no incident of aggressive behavior during or after the psychomotor automatisms
that occurred in 57 of the patients. He argued that the often-reported relationship
between aggression and psychomotor epilepsy has been exaggerated.

2 It had been frequently stated that a cancerous tumor (glioblastoma multiforme)
was situated in the amygdala. Actually, because of the mishandling of the brain at the
time of autopsy, the location of the tumor was never clearly established (Frank
Ervin, personal commun.).

phenomenon. If drug-related crimes are excluded, most of the
present upsurge in violence can be related to the rejection of
previously accepted social roles, the large numbers of people
who do not believe they have a vested interest in the stability of
our society, and the increasing belief that our institutions
cannot or will not initiate the changes that are needed. These
are not easy problems to remedy, but we will surely be in
serious trouble if a number of influential people become
convinced that violence is mainly a product of a diseased brain
rather than of a diseased society.

PSYCHOSURGERY

The current controversy over what has been called the
"resurgence of psychosurgery" places a responsibility on those
of us studying the brain and behavior— whether or not we
welcome the opportunity— to offer some light in the midst of all
this heat. Anyone who has participated in a public discussion of
this issue realizes that psychosurgery is one of those topics on
which most people prefer to have one soul-satisfying emotional
outburst rather than attempt to draw conclusions from very
complex and often conflicting data. While I have nothing
against emotional catharsis, there is an obligation to examine
the logic of the arguments and the relevant evidence as
impartially as possible, if we are to make a contribution to
something besides our own psychological well-being. Some of
the political and social arguments that have been introduced
have aroused such passion that people are forced to take sides
on these issues and in the process forget that there may be a
patient in desperate need of help. It is possible to make only a
few remarks and I offer these in an effort to set the stage for
some constructive dialogue by placing the problem in perspec-
tive. The serious ethical and legal questions concerning in-
formed consent, adequate review of experimental medical
procedures, and operations on children or those committed to
psychiatric and penal institutions cannot be discussed here (see
Shapiro, 1974, and Valenstein, 1973).

No discussion on this topic would be complete without at
least one person arguing against psychosurgery by reminding us
that the brain is the seat of our personality, humanity,
creativity, capacity to Learn, to experience emotion, and even of
our soul. 1 It is certainly true that if we remove the brain all of
these capacities will be lost, with the possible exception of the
soul. I do not want to appear facetious or to denigrate these
human qualities, but 1 want to emphasize that we must talk
about particular parts of the brain and the functions that are
regulated by these parts. It is well known that many people
have had localized brain tumors removed with little, if any,
detectable loss in these human capacities.

It is often argued that psychosurgery is unique in that
healthy tissue is destroyed for a presumed therapeutic purpose.
In truth, however, psychosurgery is really not that unique in this
regard. There are several medical procedures that involve the
destruction of healthy tissue in order to accomplish some
therapeutic advantage. For example, removal of a normal
endocrine gland to arrest some pathological process is not
uncommon. Unquestionably, there are important differences
between removing an endocrine gland, where replacement
hormonal therapy is possible, and destroying part of the brain,
but there also exist procedures other than psychosurgery that
involve destruction of normal brain tissue. It is instructive to
consider a few such examples.

Dr. Irving Cooper of St. Barnabas Hospital in New York has
done more than 10,000 brain operations on patients suffering
from such movement disorders as Parkinsonian tremors, various
types of spasticity, and choreoathetosis. While not everyone
concurs. Cooper (1969) reported a high percentage of success.

1 For example, the preamble to a bill controlling psychosurgery passed in June,
1973, by the Oregon State legislature (Senate Bill 298) reads: "Whereas it is acknowl-
edged that the human brain is the organ which gives man his unique qualities of
thought and reason, personality and behavior, emotion and communication. And,
indeed, is that unique structure importing to man his soul and ethical being; and

"Whereas these things being so, the free and full use of brain is the absolute and
inalienable right of each individual, a prerequisite for making choices, possessing
insight and judgement, and in health providing for the exercise of citizenship . . ."

In all likelihood, Cooper destroyed healthy brain tissue (in the
ventral thalamus or basal ganglia) as he freely admits in his
writings. It is important to appreciate that in many instances
there is some loss of function unrelated to the regulation of
movement that is incurred. For example, in one review Cooper
and his colleagues (Cooper et al., 1968) pointed out that
following surgery 58 percent of the patients suffer "mild," and
28 percent "moderate," deficits in speech articulation, phona-
tion, and even the selection of appropriate words. The danger of
such undesirable side effects does not necessarily rule out a
therapeutic procedure. The risks must be weighed against the
possible benefits.

To cite a different example. Many cases of temporal lobe
epilepsy are classified as idiopathic— that is, of unknown origin.
Indeed, Dr. Wilder Penfield of the Montreal Neurological
Institute wrote that he believed that in a number of instances
the basic disorder may actually exist in some subcortical region
and be projected to the temporal lobe. Nevertheless, there are
many people with excellent credentials and extensive experi-
ence who would agree that the removal of a restricted part of
the temporal lobe has helped patients with otherwise intractable
episodes of seizures, although here too undesirable side effects—
in some cases serious— are not unknown.

In other cases of intractable epilepsy the cutting of the
corpus callosum, the most extensive fiber connections between
the two sides of the brain, has significantly decreased the
incidence of seizures according to Drs. Bogen and Vogel of the
California College of Medicine. No one believes that the corpus
callosum in these patients was not perfectly normal before
surgery. Here too there were deficits produced by the surgery.
Sperry and his colleagues, for example, have demonstrated
striking deficits in these "split-brain" people, but it takes special
testing to reveal them (see Gazzaniga, 1970). Postoperatively,
the patients function quite well in normal life situations,
certainly much better than when they were plagued by a
number of grand mal seizures every day.

Admittedly these surgical procedures are controversial and

drugs have decreased the need for them. It should be noted,
however, that many would argue that these surgical techniques
are still very helpful for the elimination of some intractable
symptoms and that a loss-benefit analysis would justify their
use. Therefore, with respect to the issue of destroying healthy
tissue, psychosurgery should not be thought of as a unique
therapeutic practice. It is more realistic to view it as one end of
a continuum differing mostly on the clarity of the diagnosis
rather than the treatment.

It is true that as of now there is virtually no reliable evidence
linking psychiatric disorders to brain pathology. 1 It is important
to note, however, that there are few brain scientists prepared to
rule out the possibility that significant relationships between
psychiatric condition and brain abnormalities may be found in
the future. One of the difficulties thus far encountered in the
search for a relationship is that evidence of pathology in the
nervous system is much more subtle than it is in other organs. It
certainly is possible that functional abnormalities in the brain of
psychotic patients can never be detected by the relatively low
magnification of the light microscope. It has been reported that
the electron microscope has revealed significant defects in the
fine aborizations of neurons in the brains of some mental
defectives. It is possible that the greater degree of magnifica-
tion afforded by the electron microscope may reveal structural
abnormalities in selective regions of the brains of some
psychiatric patients.

Unless one argues for the independence of mind and body,
the possibility of structural or biochemical abnormalities cannot
be ruled out. It should be noted that even if regional brain
abnormalities are found, it is not necessary to assume that these
were the initial cause of the psychiatric disorder. Abnormal

1 Dr. Fred Plum's recent observation that "schizophrenia has been the graveyard of
many neuropathologists" refers to the fact that a large number of early pathologists
wasted much of their professional lives pursuing false leads. These leads could not be
substantiated by others or were shown to be brain artifacts resulting from the deteri-
orated physical condition of long term institutionalized patients (see Kety and
Matthysse, 1972).

brain functioning could be a by-product of abnormal behavior
produced by environmental contingencies. Nevertheless, once
produced, such brain functioning could play a major role in
maintaining abnormal behavior, emotionality, and thought
processes. We certainly do not object to this type of reasoning
when applied to disorders that we label psychosomatic. When a
substantial number of neuroscientists believe that brain abnor-
malities, perhaps of a biochemical nature, will eventually be
linked to some psychiatric disorders, measures that close the
door to future investigation of this possibility should be
discouraged. 1

Still another argument raised is that the rationale for
psychosurgery, that is, the physiological evidence that justifies
the procedure, is very primitive. This is true enough and I have
discussed the problem in detail elsewhere (Valenstein, 1973).
We should observe, however, that a number of medical
treatments is based on the empirical evidence that they work
despite the fact that understanding of the physiological mecha-
nisms responsible for their action are not available. If we
demanded a good rationale for all medical treatment we would
not even use aspirin, not to mention psychopharmacological
drugs and electroconvulsive shock treatment. (Incidentally,
despite considerable criticism of the possible overuse of
electroconvulsive treatment and its poorly understood mecha-
nisms for inducing change, the majority of psychiatrists
maintain that it is still the most effective way of arresting some
cases of very severe depression.)

Judging from accounts in the popular news media, the issue
that has caused the most concern is the charge that psychosur-

'Of interest here is a recent poll of the Society tor Neuroscience, an organization
that includes among its members most of the leading brain scientists in this country.
Of the 873 respondents, 74% (16% disagreed and 10% had no opinion) expressed the
belief that psychosurgery should be available to patients suffering from incapacitating
mental disorders provided adequate safeguards are taken. A great majority (76%) of
the members felt, however, that a commission should be established "to promulgate
guidelines for selecting and evaluating patients, for certifying that there is a recog-
nized functional disorder, for determining that psychosurgery is an appropriate last
resort, for obtaining informed consent and for follow-up and record keeping."

gery may be used as a political instrument to control people
particularly so-called militant blacks. These charges have been
accepted as true and repeated by many people who have made
no effort to check the facts. My own view, after carefully
surveying the literature and doing some direct checking, is that
the charges cannot be substantiated and that they were really
demagogic attempts to add emotional fire to the issue and to
secure political allies. 1 It is clear that we have to be vigilant and
monitor carefully the practices in state and private institutions
where there may be disproportionate representations based on
race, social class, or sex. As real and as serious as that problem
may be, however, it is quite different from some of the charges
we have been hearing. It should be noted that a substantial
proportion of the 500 to 600 psychosurgical patients operated
on in the United States each year are not institutionalized, but
are private patients referred by psychiatrists.

In the minds of many, psychosurgery is thought of as a
behavior-control technique of potentially wide applicability
rather than as an experimental therapeutic procedure for
intractable psychiatric disorders. This belief has had a very
significant influence on legislation presently being considered.
For example, in the proposed federal legislation (H.R. 6852)

1 To the best of my knowledge, the person most responsible for this belief is the
psychiatrist Peter Breggin. Breggin has charged that "these brain studies are not
oriented toward liberation of the patient. They are oriented toward law and order
and control-toward protecting society against the so-called radical individual." In his
statement attacking psychosurgery, which was read into the Congressional Record
(February 24, 1972, vol. 18, no. 26), Breggin implied that Dr. O. J. Andy, a Missis-
sippi neurosurgeon, concealed that he was operating mainly on blacks. This and
similar charges have been repeated by many people as well as in such magazines as
Ebony (Mason, 1973) apparently without troubling to check the facts. However, in
answer to my inquiry, Dr. Andy wrote that of the approximately 40 psychosurgical
operations he has performed, only 5% (i.e., 2 cases) were black. At a symposium on
psychosurgery at the 197 3 American Psychological Association Meeting in Montreal,
Dr. William Scovillc, the outgoing president of the International Psychosurgical Asso-
ciation, stated that he has never performed psychosurgery on a black person. The
speculation by Mark. Sweet, and Ervin (1967) that the more violent participants in a
riot may have some brain pathology has undoubtedly caused much anxiety about
future applications. Nevertheless, their psychosurgical patient population does not
reflect any racial bias.

outlawing psychosurgery these procedures are defined as brain
surgery for the purpose of:

"(A) modification or control of thoughts, feelings, actions, or
behavior rather than the treatment of a known and diagnosed
physical disease of the brain;

"(B) modification of normal brain function or normal brain
tissue in order to control thoughts, feelings, action, or behavior"

Similar wording can be found in other proposed legislation or
legislation that has already been passed. Clearly the concern
that these techniques will be used to control people has
provided a good part of the motivational impetus behind such
legislation. It is understandable that black congressmen and
women are among the leading supporters of the above legisla-
tion. Apparently, they have been convinced that psychosurgery
is a technique for controlling behavior that has been or is likely
to be selectively used against one segment of the population. It
is most important that precedent-setting legislation aimed at
curtailing experimental medical procedures be considered care-
fully and not hastily framed in response to a distorted
representation of the problem.

This critique of many of the common arguments against
psychosurgery should not be construed as my support for
these surgical procedures. My reasons for presenting this point
of view are twofold. On the one hand, I believe that if
psychosurgery is criticized on the wrong grounds the legislative
remedies may take a form that would establish a dangerous
precedent. Also, a criticism of irrelevant arguments or unsub-
stantiated charges can help to focus our attention on what
should be the main issue, namely, Can destruction of a part of
the brain be justified on therapeutic grounds? This question is
easier to ask than to answer. Even if all the data on the
consequences of a particular psychosurgical procedure were in
agreement and their meaning unambiguous, it would still be
possible to reach opposite conclusions because of personal
weights assigned to gains and losses in different capacities. Is a
flattening of emotional responsiveness, for example, balanced
by freedom from a crippling anxiety?

It is not possible for me to present any firm eonclusions, let
alone to substantiate them, on the approximately one-dozen
different brain operations that eould be called psychosurgery.
Raising some of the main problems that will have to be faced in
evaluating any psychosurgical procedure may serve some useful
purpose. To begin, we have to face the likelihood that the
results of any brain operation probably will always contain an
element of unpredictability that will not be completely elimi-
nated by any increased technical precision. This is true in part
because the ramifications of destroying any part of the brain
must depend upon the total personality of the patient, or if you
prefer, on the total neuronal context that must mediate the
impact of destruction of any one part of the brain. Moreover,
there is usually some compensation for loss in function
following brain damage, but the amount of compensation varies
with individuals for a great number of reasons we cannot go
into at this time.

Another problem in evaluating psychosurgery is that the
available evidence leaves much to be desired. In the first place,
most of the testing of patients following psychosurgery was
done at a time when the patient population and the surgical
procedures were different from those that exist today. The
older prefrontal lobotomy procedures destroyed much larger
brain areas than do the current so-called fractional operations.
Although most of the older operations involved rotating surgical
knives inside the brain in order to disconnect large areas of the
prefrontal cortex, present-day techniques may limit destruction
to an area 3 to 5 mm. in diameter. There is also little doubt that
the more modern methods of stereotaxic surgery make it
possible to reach specific brain targets with much more
precision than was previously possible.

No purpose is served by reviewing in detail the results of the
older prefrontal lobotomy procedures. The results were ex-
tremely variable and one can without difficulty find evidence
on both sides of the controversy. There is evidence in the
literature demonstrating a blunting of emotional responsiveness,
lowering of performance on at least some parts of IQ tests, an

inability to maintain goal-directed behavior, the triggering of
epileptic seizures, and other neurological problems following
prefrontal lobotomy. There are also a number of studies that
reported significant psychiatric improvement following the
operations, no IQ loss, and an increased ability to hold a job.
Some of the studies that reached this positive conclusion
involved relatively long-term follow-ups and some, such as those
conducted by the Connecticut Lobotomy Committee or the
British Board of Control Study, included substantial samples of
patients (Moore et al., 1948). The Columbia-Greystone study,
which involved more than 50 participating investigators and a
battery of 35 psychological tests (selected from a list of more
than 100 that were considered), concluded that there was no
evidence that topectomy (one type of prefrontal operation)
produced any permanent loss in learning ability, memory,
creativity, imagination, intellectual achievement, social or ethi-
cal attitudes, or even sense of humor (see Mettler, 1949, 1952;
Landis, Zubin, and Mettler, 1950). These studies can all be
criticized on various methodological grounds; the test instru-
ments were probably insensitive to important changes in
behavioral capacities, and the estimates of improvement often
gave exaggerated weighting to the elimination of behavior
troublesome to the hospital staff or society in general while
placing considerably less emphasis on the qualitative aspects of
the postoperative adjustment level.

While we can learn much from examining the older prefrontal
lobotomy literature— particularly in respect to methodological
points in the way such studies should or should not be
conducted— it is not possible to apply specific conclusions to
the brain operations performed today. Very different brain
areas are often involved, even where the surgery is still directed
at prefrontal areas. There are fewer studies reporting results
following selective damage to limbic and hypothalamic struc-
tures. It is probably safe to conclude that the added precision of
the newer operations has resulted in many fewer instances of
gross behavioral deterioration, or neurological side-effects such
as epilepsy. However, our information about the emotional and

intellectual changes produced by the newer psychosurgical
procedures is very inadequate.

Neurosurgeons have neither the training nor the time to
conduct the type of studies needed to evaluate adequately the
changes produced by their brain operations. Postoperative
changes are usually reported in gross terms listing percentages
of patients exhibiting different degrees of improvement in
poorly defined categories ranging from "completely cured"
to "no change." There are few examples where postoperative
evaluative tests were designed to measure changes in those
capacities that animal studies have emphasized as likely to be
altered. Indeed, many neurologists and neurosurgeons have
displayed an amazing "tunnel vision" toward animal studies.
They have been quick to see clinical applications in animal
studies, but often quite blind to the results that should have
cautioned them against the operation and influenced their
evaluative procedures. A few examples are offered to illustrate
this point.

There is some familiarity with the circumstances that
encouraged Egas Moniz, the Portuguese neurologist and Nobel
laureate, to initiate prefrontal lobotomy. It will be recalled that
at the International Neurology Congress in London in 1935
Carlyle Jacobsen presented his results on the behavior changes
in chimpanzees following destruction of their frontal lobes.
Prior to the operation, one of the chimpanzees— the now-
famous Becky— had a temper tantrum every time she made a
mistake in the testing situation. After frontal lobe surgery,
however, she showed no evidence of emotional disturbance
under similar circumstances. Moniz was sitting in the audience,
and according to John Fulton, the session chairman: "Dr.
Moniz arose and asked if frontal lobe removal prevents the
development of experimental neurosis in animals and elimi-
nates frustrational behavior, why would it not be feasible to
relieve anxiety states in man by surgical means." The main
thrust of Jacobsen's presentation, namely, that the operated
animals were no longer able to perform certain problem-solving
tasks (particularly those involving delayed responses) was

ignored. Within three months, Moniz had persuaded his neuro-
surgical colleague, Almeida Lima, to operate on their first
patient.

Anterior cingulotomy is another psychosurgical procedure
used by several surgeons today. Here, too, a careful reconstruc-
tion of the history reveals a striking "tunnel vision." John
Fulton's description of the animal experiments by Wilbur Smith
(1945) and Arthur Ward (1948) in a number of influential
speeches had a direct influence on the adoption of cingulotomy
procedures by a number of people in England, France, and in
the United States. Fulton reported that following cingulotomy
monkeys became tamer. A closer examination of Ward's
description of the postoperative behavior of the monkeys
reveals the inadequacy of the term "tameness" to summarize all
the changes that occurred. For example, Ward said:

there is an obvious change in personality. The monkey loses its
preoperative shyness and is less fearful of man. It appears more
inquisitive than the normal monkey of the same age. In a large
cage with other monkeys of the same size, such an animal shows no
grooming behavior or acts of affection towards its companions. In fact,
it treats them as it treats inanimate objects and will walk on them,
bump into them if they happen to be in the way, and will even sit on
them. It will openly eat food in the hand of a companion without being
prepared to do battle and appears surprised when it is rebuffed. Such an
animal never shows actual hostility to its fellows. It neither fights nor
tries to escape when removed from a cage. It acts under all
circumstances as though it had lost its "social conscience." This is
probably what Smith saw and called "tameness." It is thus evident that
following removal of the anterior limbic area, such monkeys lose some
of the social fear and anxiety which normally governs their activity
and thus lose the ability to accurately forecast the social repercussions
of their own actions.

Perhaps the most striking example of "tunnel vision" comes
from a psychosurgical procedure that involves destruction of
the ventromedial hypothalamus in persons diagnosed as pedo-
philic homosexuals, that is, men who seek out sexual opportuni-
ties with young boys. Dr. F. Roeder and his colleagues at the
University of Gottingen in Germany received their inspiration
while watching a film at another International Neurology
Congress, held in Brussels in 1957 (Roeder et al., 1971, 1972).

Roedcr described his response to this film which depicted the
hypersexual behavior of cats amygdalectomized by Leon
Schreiner and Arthur Kling. "the behavior of male cats with
lesions of the amygdalar region in some respects closely
approached that of human perversion. The films convinced us
that there was a basis for a therapeutic stereotaxic approach to
this problem in man. 1 ' Roeder was referring to work on cats by
Arthur Kling which demonstrated that ventromedial hypotha-
lamic lesions eliminated the hypersexuality previously produced
by amygdala lesions. Roeder and his colleagues proceeded to
make stereotaxic lesions in the ventromedial hypothalamic
nucleus in man. Based on experience with a relatively small
patient population studied in a cursory way, Roeder and his
associates reached the disquieting if not shocking, conclusion
about their surgical procedure that "there is no doubt that
experimental behavioral research has afforded us a basic method
to eliminate or to control pedophilic homosexuality by means
of an effective psychosurgical operation in the area of the sex
behavior center." Those of us who study the brain and behavior
in animals know of the voluminous literature implicating the
ventromedial hypothalamic nucleus in endocrine regulation,
appetite, and many other functions. There is also good evidence
that irritability and aggressiveness can be produced by lesions in
this area. However, once the focus was directed at sexual
behavior, the other important behaviors regulated by this brain
area were ignored.

Similar comments could be made in reference to a recent
report on producing stereotaxic lateral hypothalamic lesions to
combat obesity in humans (Quaade, 1974). As Marshall (1974)
pointed out in a comment on Quaade's report, the lateral
hypothalamus is not specifically involved in "monitoring the
energy needs of the organism and transforming such informa-
tion into an urge to eat." In animals, lateral hypothalamic
damage also produces sensory changes leading to inattentiveness
to external stimuli and impairment in sexual activation, learning
ability, and memory.

A point that apparently has to be made over and over again is

that there are very few parts of the brain that control only one
behavior. People studying a given area of the brain may
emphasize either control of appetite, aggression, endocrine
balance, or sexual behavior, and so forth, depending on their
own interests. I have stressed this "tunnel vision" problem
because it illustrates the danger of superficial contacts between
experimentalists and clinicians. There are many consequences of
this lack of communication. Obviously, in some instances,
operations should never have been performed. In a great many
instances, behaviors and capacities that should have been
assessed were completely neglected in the postoperative eval-
uation of patients. What is needed is not some hastily conceived
legislation that may set a precedent hindering all investigations
in experimental medicine. We clearly need better controls to
protect patients, but it must be recognized that this cannot be
accomplished unless more meaningful interactions between
research scientists and clinicians are established.

LITERATURE CITED

Adams, 1) . and J P. Nynn
1966. Transfer of an escape response from tail shock to brain stimulated attack
behavior. Jour. Exp. Anal. Behavior, vol. 9, pp. 401-408.

Andersen. R.
1972. Differences in the course of learning as measured by various memory tasks
alter amygdalectomy in man. //; Hitchcock, E., L. Laitinen, and K. Vaernet
(eds.). Psychosurgery. Springfield, 111., Charles C. Thomas, pp. 177-183.

Balasubramaniam. V., T. S. Kanaka, and B. Ramamurthi
1970. Surgical treatment of hyperkinetic and behavior disorders. Int. Surg., vol.
54, pp. 18-23.
Ball, G. G.
1974. Vagotomy: effect on electrically elicited eating and self-stimulation in the
lateral hypothalamus. Science, vol. 184, pp. 484-485.

Black. S. L., and C. H. Vanderwolf
1969. Thumping behavior in the rabbit. Physiol. Behavior, vol. 4, pp. 445-449.

Bowden, D. M., T. Galkin, and H. E. Rosvold
[In press.) Plasticity of the drinking system as defined by electrical stimulation of
the brain (ESB) in monkeys. Physiol. Behavior.

Brown, S., and E. A. Schafer
1888. An investigation into the functions of the occipital and temporal lobes of
the monkey's brain. Phil. Tr.,vol. 179B, pp. 303-327.

Burckhardt, G.
1 89 1 . Ucber Rindenexcisionen, als Bertrag zur operativen Therapie der Psychosen.
Allg. Psychiat., vol. 47, pp. 463-548.

Clark, A. C.
1964. Profiles of the future. New York, Bantam.

Cooper, I. S.

1969. Involuntary movement disorders. New York, Harper and Row, Hoeber Med.
Div.

Cooper, 1. S., M. Riklan, S. Stellar, J. M. Waltz, E. Levita, V. A. Ribera, and J.
Zimmerman

1968. A multidisciplinary investigation of neurosurgical rehabilitation in bilateral
Parkinsonism. Amer. Geriat. Soc, vol. 16, pp. 1 177-1 306.

Cox, V. C, and E. S. Valenstein

1969. Distribution of hypothalamic sites yielding stimulus-bound behavior. Brain,
Behavior, Evol., vol. 2, pp. 359-376.

Crichton, M.
1972. The terminal man. New York, Alfred Knopf.

Deadwyler, S. A., and E. J. Wyers
1972. Description of habituation by caudate nuclear stimulation in the rat.
Behavioral Biol., vol. 7, pp. 55-64.

Delgado, J. M. R.
1969. Physical control of the mind. New York, Harper and Row.

Divac, I., H. E. Roswold, and M. K. Szwarcbart

1967. Behavioral effects of selective ablation of the caudate nucleus. Jour. Comp.
Physiol. Psychol., vol. 63, pp. 184-190.

Doty, R. W.

1969. Electrical stimulation of the brain in behavioral context. Ann. Rev.
Psychol., vol. 20, pp. 289-320.

Ellison, G. D., and J. P. Flynn

1968. Organized aggressive behavior in cats after surgical isolation of the
hypothalamus. Arch. Italiennes Biol., vol. 106, pp. 1-20.

Ferrier, D.
1876. The Functions of the Brain. London, Smith Elder & Co.

Flynn, J. P., S. B. Edwards, and R. J. Bandler
1971. Changes in sensory and motor systems during centrally elicited attack.
Behavioral Sci., vol. 16, pp. 1-19.

Folkow, B., and E. H. Rubinstein
1965. Behavioral and autonomic patterns evoked by stimulation of the lateral
hypothalamic area in the cat. Acta Physiol. Scandinavica, vol. 65, pp.
292-299.

Fritsch, G., and E. Hitzig
1870. Ueber die elektrische Erregbarkeit des Grosshirns. Arch. Anat. Physiol.,
Leipzig, vol. 37, pp. 300-332.

Gazzaniga, M. S.

1970. The bisected brain. New York, Appleton-Century-Crofts.

Gloor, P.
1967. Discussion. In Clemente, C. D., and D. B. Lindsley (eds.), Aggression and
defense, neural mechanisms and social patterns. Los Angeles, University of
California Press, pp. 116-124.

Goltz, F.
1892. Der -Hund ohne Grosshirn. Pfliiger's Arch. Ges. Physiol., vol. 51, pp.
570-614.

Greer, M. A.
1955. Suggestive evidence of a primary "drinking center" in the hypothalamus of
the rat. Proc. Soc. Exp. Biol. Med., vol. 89, pp. 59-62.

Heath, R. G., S. B. John, and C. J. FontUa

1968. The pleasure response: studies by stereotaxic technique in patients. In
Kline, N., and E. Laska (eds.). Computers and electronic devices in
psychiatry. New York, Grune and Stratton, pp. 178-189.

Heimburger. R. F., C. C. Whitlock, and J. E. Kalsbeck
1966. Stereotaxic amygdalectomy for epilepsy with aggressive behavior. Jour.
Amer. Med. Assoc, vol. 198, pp. 741-745.

Hess. W. R
1957. The functional organization of the diencephalon. New York, Grune and
Stratton.

Ingraham, B. L.. and G. W. Smith
1972. The use of electronics in the observation and control of human behavior
and its possible use in rehabilitation and parole. Issues in Criminology, vol.
7, pp. 35-53.

Kellaway, P.
1946. The part played by electric fish in the early history of bioelectricity and
electrotherapy. The William Osier Medical Essay. Bull. Hist. Med., vol. 20,
pp. 112-137.

Kety, S.S., and S. Matthysse (eds.)

1972. Prospects for research on schizophrenia. Neurosci. Res. Program Bull., vol.
10, no. 4, pp. 384-388.

Kim, Y. K., and W. Umbach

1973. Combined stereotaxic lesions for treatment of behavior disorders and severe
pain. In Laitinen, L. V., and K. E. Livingston (eds.), Surgical approaches in
psychiatry. Baltimore. Md., University Park Press, pp. 182-188.

Kirkby. R. J. and D. P. Kimble
1968. Avoidance and escape behavior following striatal lesions in the rat. Exp.
Neurol., vol. 20, pp. 215-227.

Kling, A.
1972. Effects of amygdalectomy on social-affective behavior in non-human
primates. In Eleftheriou, B. E. (ed.), The neurobiology of the amygdala.
New York. Plenum Press, pp. 51 1-5 36.

Kling, A., J. Lancaster, and J. Benitone
1970. Amygdalectomy in the free ranging vervet (Cercopithecus althiops). Jour.
Psychiat. Res., vol. 7, pp. 191-199.

Kliiver, H., and P. C. Bucy
1939. Preliminary analysis of functions of the temporal lobe in monkeys. Arch.
Neurol. Psychiat. (Chicago), vol. 42, pp. 979-1000.

Konig, J. F. R., and R. A. Kippel

1963. The rat brain: A stereotaxic atlas of the forebrain and lower parts of the
brain stem. Baltimore, Md., William and Wilkins.

Kubie, L. S.
1953. Some implications for psychoanalysis of modern concepts of the organiza-
tion of the brain. Psychoanal. Quart., vol. 22, pp. 21-52.

Landis, C, J. Zubin, and F ; . Mettler
1950. The functions of the human frontal lobe. Jour. Psychol., vol. 30, pp. 123-138.

Leyton, A. S. F., and C. S. Sherrington
1917. Observations on the excitable cortex of the chimpanzee, orang-utan, and
gorilla. Quart. Jour. Exp. Physiol., vol. 11, pp. 135-222.

Libet, B.
1973. Electrical stimulation of cortex in human subjects and conscious sensory
aspects. In Iggo, A. (ed.). Handbook of sensory physiology, vol. 2,
somatosensory system. Berlin, Springer- Verlag, pp. 743-790.

London, P.

1969. Behavior control. New York, Harper and Row.

Luria, A. R.

1973. The working brain. London, Penguin Press.

Mahl, G. F., A. Rothenberg, J. M. R. Delgado, and H. Hamlin

1964. Psychological responses in the human to intracerebral electric stimulation.
Psychosom. Med., vol. 26, pp. 337-368.

Mark, V. H., and F. R. Ervin

1970. Violence and the brain. New York, Harper and Row.

Mark, V. H., W. H. Sweet, and F. R. Ervin
1967. Role of brain disease in riots and urban violence (letter to the journal).
Amer. Med. Assoc, vol. 201, p. 895.

Marshall, J. F.

1974. Stereotaxy for obesity. Lancet, vol. 1, no. 7872, p. 106.

Mason, B. J.
1973. New threat to Blacks: brain surgery to control behavior. Ebony, vol. 28, pp.
63-72.

Mettler, F. A. (ed.)
1949. Selective partial ablation of the frontal cortex. A correlative study of its
effects on human psychotic subjects. New York, Paul Hoeber.

1952. Psychosurgical problems (The Columbia Grey stone Associates, Second

Croup). New York, BlakistOIl

Mettler, F. A., and C. Mettles
1942. The effects of striatal injury Brain, vol. 65, pp. 242-255.

Miller, N. E.
1973. Commentary. In Valenstein, E S. (ed). Brain stimulation and motivation.
Glenview, 111., Scott, Poresman, pp. 53-68.

Moore, B., S. Friedman, B. Simon, and J. Parmer

1948. Connecticut Lobotomy Committee: a cooperative clinical study of lobot-
omy. Res. Publ. Assoc. Nerv. Ment. Dis., vol. 27, pp. 769-794.

Niven, L.

1970. Death by ecstasy. In Wolheim, D. A., and T. Carr (eds.). World's best
science fiction. New York, Ace.

Olds, J.
1973. Commentary: In Valenstein, E. S. (ed.), Brain stimulation and motivation.
Glenview, 111., Scott, Foresman, pp. 81-99.

Olds, J., and P. Milner
1954. Positive reinforcement produced by electrical stimulation of septal area and
other regions of the rat brain. Jour. Comp. Physiol. Psychol., vol. 47, pp.
419-427.

Panksepp, J.

1971. Aggression elicited by electrical stimulation of the hypothalamus in albino
rats. Physiol. Behavior, vol. 6, pp. 321-329.

Paxinos, G, and D. Bindra

1972. Hypothalamic knife cuts: effects on eating, drinking, irritability, aggression
and copulation in the male rat. Jour. Comp. Physiol. Psychol., vol. 79, pp.
219-229.

Penfield, W., and E Boldrey

1937. Somatic motor and sensory representation in the cerebral cortex of man as
studied by electrical stimulation. Brain, vol. 60, pp. 389-443.

Penfield, W., and P. Perot

1963. The brain's record of auditory and visual experience: A final summary and
discussion. Brain, vol. 86, pp. 595-696.

Penfield, W., and T. Rasmussen
1950. The cerebral cortex of man. New York, Macmillan.

Penfield, W., and K. Welch

1949. Instability of response to stimulation of the sensori-motor cortex of man.
Jour. Physiol. (London), vol. 109, pp. 358-365.

Phillips, A. G., V. C. Cox, J. W. Kakolewski, and E. S. Valenstein

1969. Object-carrying by rats: an approach to the behavior produced by brain
stimulation. Science, vol. 166, pp. 903-905.

Plotnik, R.
1974. Brain stimulation and aggression: monkeys, apes, and humans. In Holloway,
R. L. (ed.), Primate aggression, territoriality and xenophobia: a comparative
approach. New York, Academic Press, pp. 138-149.

Plotnik, R., and J. M. R. Delgado

1970. Emotional responses in monkeys inhibited with electrical stimulation.
Psychonomic Sci., vol. 18, pp. 129-130.

Quaade, F.
1974. Stereotaxy for obesity. Lancet, vol. 1, no. 7851, p. 267.

Reis, D. J., N. Doba, and M. A. Nathan
1973. Predatory attack, grooming, and consummatory behaviors evoked by
electrical stimulation of cat cerebellar nuclei. Science, voL 182, pp.
845-847.

Robinson, B. W., M. Alexander, and G. Browne
1969. Dominance reversal resulting from aggressive responses evoked by brain
telestimulation. Physiol. Behavior, vol. 4, pp. 749-752.

Rodin, E. A.
1973. Psychomotor epilepsy and aggressive behavior. Arch. Gen. Psychiat., vol.
28, pp. 210-213.

Roeder, F., and D. Miiller
1969. The stereotaxic treatment of paedophillic homosexuality. Stuttgart, Ger-
man Med. Monthly (G. R. Graham, ed. English ed.). Georg Thime Verlag,
vol. 14, pp. 265-271.

Roeder, F., D. Miiller, and H. Orthner

1971. Stereotaxic treatment of psychoses and neuroses. In Umbach, W. (ed.),
Special topics in stereotaxis. Stuttgart, Hippokrates-Verlag, pp. 82-105.

Roeder, F., H. Orthner, and D. Miiller

1972. The stereotaxic treatment of pedophilic homosexuality and other sexual
deviations. In Hitchcock, E., L. Laitinen, and K. Vaernet (eds.), Psycho-
surgery. Springfield, 111., Charles C. Thomas, pp. 87-1 1 1.

Rorvik, D.
1969. Someone to watch over you (for less than 2 cents a day). Esquire, voL 72,
p. 164.

Rosvold, H. E., A. F. Mirsky, and K. H. Pribram
1954. Influence of amygdalectomy on social behavior in monkeys. Jour. Comp.
Physiol. Psychol., vol. 47, pp. 173-178.

Rosvold, H. E., M. Mishkin. and M. K. Szwarcbart
1958. Effects of subcortical lesions in monkeys on visual-discrimination and
single-alternation performance. Jour. Comp. Physiol. Psychol., vol. 51, pp.
437-444.
Routtenberg, A., and J. Lindy
1965. Effects ol the availability of rewarding septal and hypothalamic stimulation
on bar-pressing for food under conditions of deprivation. Jour. Comp.
Physiol. Psychol., vol. 60, pp. 158-161.

Schwartzbaum, J.
I960. Changes in reinforcing properties of stimuli following ablation of the
amygdaloid complex in monkeys. Jour. Comp. Physiol. Psychol., vol. 53,
pp. 388-395.

Shapiro, M. H.
1974. Legislating the control of behavior control: autonomy and the coercive use
of organic therapies. Southern California Law Rev., vol. 47, pp. 237-356.

Smith, W. K.
1945. The functional significance of the rostral cingular cortex as revealed by its
response to electrical stimulation. Jour. Neurophysiol., vol. 8, pp. 241-255.

Stein, L.
1971. Neurochemistry of reward and punishment: some implications for the
etiology of schizophrenia. Jour. Psychiat. Res., vol. 8, pp. 345-361.

Stevens, J. R., V. H. Mark, F. Ervin, P. Pacheco, and K. Suematsu
1969. Deep temporal stimulation in man: long latency, long lasting psychological
changes. Arch. Neurol. (Chicago), vol. 21, pp. 157-169.

Valenstein, E. S.

1969. Behavior elicited by hypothalamic stimulation. A prepotency hypothesis.
Brain, Behavior, Evol., vol. 2, pp. 295-316.

1971. Channeling of responses elicited by hypothalamic stimulation. Jour.
Psychiat. Res., vol. 8, pp. 335-344.

1973. Brain control: a critical examination of brain stimulation and psycho-
surgery. New York, John Wiley.

Valenstein, E. S., and B. Beer
1964. Continuous opportunity for reinforcing brain stimulation. Jour. Exp. Anal.
Behavior, vol. 7, pp. 1 83-184.

Valenstein, E. S., V. C. Cox, and J. W. Kakolewski

1968a. Modification of motivated behavior elicited by electrical stimulation of the
hypothalamus. Science, vol. 159, pp. 1119-1121.

1968b. The motivation underlying eating elicited by lateral hypothalamic stimula-
tion. Physiol. Behavior, vol. 3, pp. 969-971.

1970. Reexamination of the role of the hypothalamus in motivation. Psychol.
Rev., vol. 77, pp. 16-31.

Valenstein, E. S., J. W. Kakolewski, and V. C. Cox

1968. A comparison of stimulus-bound drinking and drinking induced by water
deprivation. Communications Behavior Biol., pt. A, vol. 2, pp. 227-233.

Valenstein, E. S., and T. Valenstein
1964. Interaction of positive and negative reinforcing neural systems. Science, vol.
145, pp. 1456-1458.

Van Buren, J. M.

1961. Sensory motor and autonomic effects of mesial temporal stimulation in
man. Jour. Neurosurg., vol. 18, pp. 273-288.

1963. Confusion and disturbance of speech from stimulation in the vicinity of the
head of the caudate nucleus. Ibid., vol. 20, pp. 148-157.

1966. Evidence regarding a more precise localization of the posterior frontal-
caudate arrest response in man. Ibid., vol. 24, pp. 416-417.

Ward, A. A., Jr.
1948. The anterior cingular gyrus and personality. Res. Publ. Assoc. Nerv. Ment.
Dis., vol. 27, pp. 438-445.

Wasman, M., and J. P. Flynn

1962. Directed attack elicited from hypothalamus. Arch. Neurol., vol. 6, pp.
220-227.

Wise, R. A.

1971. Individual differences in effects of hypothalamic stimulation: the role of
stimulation locus. Physiol. Behavior, vol. 6, pp. 569-572.

£0 I .^

.46

* ^V^C^^^*>%^pi

i" i -LJji

FORTY-SIXTH
1 JAMES ^.RTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

fTH

WHAT SQUIDS AND OCTOPUSES

TELL US
ABOUT BRAINS AND MEMORIES

JOHN Z. YOUNG

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1977

fffc

Museum of a,

1869
THE LIBRARY

FORTY-SIXTH

JAMES ARTHUR LECTURE ON

THE EVOLUTION OF THE HUMAN BRAIN

FORTY-SIXTH

JAM IS ARTHUR LECTURE ON

I III EVOLUTION OF 1 111 HUMAN BRAIN

WHAT SQUIDS AND OCTOPUSES

TELL US
ABOUT BRAINS AND MEMORIES

JOHN Z. YOUNG

Professor Emeritus and Honorary Fellow
University College, London

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1977

ADD , k m-?-7

JAMES ARTHUR LECTURES ON
THE EVOLUTION OF THE HUMAN BRAIN

Frederick Tilney, The Brain in Relation to Behavior; March 15, 1932

C. Judson Herrick, Brains as Instruments o) f Biological Values; April 6, 1933

D. M. S. Watson, Hie Story of Fossil Brains from Fish to Man; April 24, 1934

C. U. Ariens Kappers, Structural Principles in the Nervous System; The Develop-
ment of the Forebrain in Animals and Prehistoric Human Races; April 25, 1935

Samuel T. Orton, The Language Area of the Human Brain and Some of its
Disorders; May 15, 1936

R. W. Gerard, Dynamic Neural Patterns; April 15, 1937

Franz Weidenreich, The Phylogenetic Development of the Hominid Brain and its
Connection with the Transformation of the Skull; May 5, 1938

G. Kingsley Noble, The Neural Basis of Social Behavior of Vertebrates; May 11,
1939

John F. Fulton, A Functional Approach to the Evolution of the Primate Brain; May
2, 1940

Frank A. Beach, Central Nervous Mechanisms Involved in the Reproductive
Behavior of Vertebrates; May 8, 1 94 1

George Pinkley, A History of the Human Brain; May 14, 1942

James W. Papez, Ancient Landmarks of the Human Brain and Their Origin; May 27,
1943

James Howard McGregor, The Brain of Primates; May 11, 1944

K. S. Lashley, Neural Correlates of Intellect; April 30, 1945

Warren S. McCulloch, Finality and Form in Nervous Activity; May 2, 1946

S. R. Detwiler, Structure-Function Correlations in the Developing Nervous System
as Studied by Experimental Methods; May 8, 1947

Tilly Edinger, The Evolution of the Brain; May 20, 1948

Donald O. Hebb, Evolution of Thought and Emotion; April 20, 1949

Ward Campbell Halstead, Brain and Intelligence; April 26, 1950

Harry F. Harlow, The Brain and Learned Behavior; May 10, 1951

Clinton N. Woolsey, Sensory and Motor Systems of the Cerebral Cortex; May 7,
1952

Alfred S. Romer, Brain Evolution in the Light of Vertebrate History; May 21, 1953

Horace W. Magoun, Regulatory F"unctions of the Brain Stem; May 5, 1954

*Fred A. Mettler, Culture and the Structural Evolution of the Neural System; April
21, 1955

*Pinckney J. Harman, Pale one urologic, Neoneurologic, and Ontogenetic Aspects of
Brain Phytogeny; April 26, 1956

* Davenport Hooker, Evidence of Prenatal function of the Central Nervous System in

Man; April 25, 1957

*David P. C. Lloyd, The Discrete and the Diffuse in Nervous Action; May 8, 1958

*Charles R. Noback, Die Heritage of the Human Brain; May 6, 1959

*Ernst Scharrer, Brain function and the Evolution of Cerebral Vascularization; May
26, 1960

Paul I. Yakovlev, Brain, Body and Behavior. Stereodynamic Organization of the
Brain and of the Motility -Experience in Man Envisaged as a Biological Action
System; May 16, 1961

H. K. Hartline, Principles of Neural Interaction in the Retina; May 29, 1962

Harry Grundfest, Specialization and Evolution of Bioelectric Activity; May 28, 1963

*Roger W. Sperry, Problems Outstanding in the Evolution of Brain function; June 3,
1964

*Jose M. R. Delgado, Evolution of Physical Control of the Brain; May 6, 1965

Seymour S. Kety, Adaptive functions and the Biochemistry of the Brain; May 19,
1966

Dominick P. Purpura, Ontogenesis of Neuronal Organizations in the Mammalian
Brain; May 25, 1967

*Kenneth D. Roeder, Three Views of the Nervous System; April 2, 1968

t Phillip V. Tobias, Some Aspects of the fossil Evidence on the Evolution of the
Hominid Brain; April 2, 1969

*Karl H. Pribram, What Makes Man Human; April 23, 1970

VValle J. H. Nauta, A New View of the Evolution of the Cerebral Cortex of
Mammals; May 5, 1971

David H. Hubel, Organization of the Monkey Visual Cortex; May 11, 1972

Janos Szentagothai, The World of Nene Nets; January 16, 1973

* Ralph L. Holloway, The Role of Human Social Behavior in the Evolution of the

Brain;Mny 1, 1973

*Elliot S. Valenstein, Persistent Problems in the Physical Control of the Brain; May
16, 1974

^Marcel Kinsbourne, Development and Evolution of the Neural Basis of Language;
April 10, 1975

*John Z. Young, What Squids and Octopuses Tell Us About Brains and Memories,
May 13,1976

^Unpublished.

*Published versions of these lectures can be obtained from The American Museum of
Natural History, Central Park West at 79th St., New York, N. Y. 10024.

t Published version: The Brain in Hominid Evolution, New York: Columbia Univer-
sity Press, 1971.

WHAT SQUIDS AND OCTOPUSES TELL
US ABOUT BRAINS AND MEMORIES

NEW TECHNIQUES FOR STUDIES OF THE BRAIN

To reach a better understanding of the human brain we need
to develop new ways of thinking and talking about the nervous
system in general. All our knowledge of nerve fibers and their
synapses proves to be something of a disappointment when we
try to explain complex forms of behavior, such as that of man. I
have believed for many years that to overcome this difficulty
we must try to describe as fully as possible the behavior pat-
terns and the whole nervous system. When I began research, I
thought that it might be possible to do this for lampreys and
after making some studies went so far as to write what would
now be called a research program with this in view. But on fur-
ther consideration, I decided that both the behavior and struc-
ture of the brain of these animals were too difficult to study,
mainly for technical reasons. Moreover in 1929, for the first
time, I became acquainted with octopuses and squids and quite
soon decided that their nervous systems seemed likely to pro-
vide sufficient complexity to be interesting and sufficient ac-
cessibility for anatomical study and experiment. It is not too
much to make the claim that this hope was well founded, as
we now have some understanding of all parts of the cephalopod
nervous system. We also have a lot of information about their
behavioral capacities— at least in the laboratory; less, unfortu-
nately, in their native state in the sea.

It may seem to be a vain and unjustified claim that we un-
derstand cephalopod brains so well. Of course, there is an im-
mense amount that we should like to know. But I hope that
the effort to substantiate this claim may serve to bring out
both the extent and the limitations of our knowledge of all
brains, including that of man. It may show how what we mean
by "understanding the brain" has changed over the last 50

years since this research began. This may prove to be quite a
useful exercise not only in the history of neuroscience but in
the study of the relations of science and technology in general.
We can recognize four major changes of scientific method
and capabilities since 1929 that have especially influenced neu-
rology.

1. Reliable methods of recording small changes in electrical ac-

tivity have become widely available. With these we can
follow events in nerves and brains with a very high degree
of resolution in time. Resolution in space can also be pre-
cise, but is limited to a few places in the brain at a time.

2. Electronmicroscopy has provided us with the power to study

the structure and organization of neurons with a very high
degree of resolution in space. This, unfortunately, is pos-
sible only by accepting very poor resolution in time. We
cannot follow changes from moment to moment with the
electron microscope.

3. Chromatography provides us with the power to study the mi-

crochemical composition of tissues, estimating quantities
of substances that are present in very small amounts,
though again with rather poor resolution in both space and
time. Fluorescence microscopy has also been particularly
helpful in the study of the nervous system because of its
capacity to reveal selectively the course of tracts contain-
ing biologically active amines.

4. Finally, during this period mankind has enormously enlarged

his mathematical powers of computation. Computers help
us to bring together the vast masses of data provided by
other techniques. Besides their help with arithmetical
operations, it is even more important that computers have
led to great advances in our understanding of the operations
of communication and control which, until recently, were
considered only by using the language of subjective psy-
chology.
Knowledge of the nervous system has profited from these ad-
vances. My own detailed contributions have mostly been in

humbler fields, using older techniques of histology and psy-
chology. But through developments that we have sponsored in
the Department of Anatomy at University College, London, I
have been near the beginning of several of these four major new
developments of technique and have been able to find helpers
in applying them to cephalopods.

THE BRAIN AS A HIERARCHICAL SOMATOTOPIC COMPUTER

Our aim is to try to understand the nervous system as a
whole. Let us therefore begin with the last of the new techniques
mentioned. Cybernetics can tell us how to think of the brain as
a hierarchical computer, somatotopically organized (Arbib,
1972). The idea of hierarchy in the nervous system was intro-
duced by the clinician Hughlings Jackson long ago, and cyber-
netic analysis shows that it is really an essential feature of any
organization that uses much information to accomplish a pur-
pose, whether it be an army or an octopus. Hierarchy allows
each level to receive only that part of the information that is
relevant for the decisions it must take. This is magnificently
illustrated by octopuses (fig. 1). Each of the eight arms carries
hundreds of highly mobile suckers and the movements of

■ }

FIG. 1. An octopus swimming forward to attack a crab.

these, and of the whole arm, are controlled by nerve cells lying
in ganglia within the arm. There are altogether 350 million cells
in the arms as compared with only 150 million in all the rest of
the nervous system (Young, 1971). The suckers are the enlisted
men of the cerebral army, and their local nerve cells are the
noncommissioned officers. Individual isolated arms are capable
of quite complicated coordinated movements, for example
acting either to draw objects in or to reject them. These periph-
eral centers are thus the next layer of members of the hier-
archy and can act independently. They are the regiments of
the cerebral army, and the nerve cells placed along the center of
each arm are the junior officers who control them. They receive
information from individual suckers and order them to act in
particular sequences.

The brain contains lower motor centers, comparable with our
own spinal cord (fig. 2), and these control movements of all
the arms when working together and of the mantle, which acts
by jet propulsion. Electrical stimulation of these centers will
produce movements of the relevant parts, including changes of
color by the chromatophores (Boycott and Young 1950; Boy-
cott, 1961). To pursue our analogy we here have regimental and
brigade headquarters. They receive relevant information from
the arms and send orders to them. However, these centers nor-
mally operate under the control of still higher motor centers in
the basal supraoesophageal lobes. These basal lobes have struc-
ture strikingly like our own cerebellum, but before we can un-
derstand their working we must begin to think more carefully
about what tasks the nervous system has to do, and what we
mean when we say that it sends information, instructions, or
commands.

COMMUNICATION AND CONTROL BY THE NERVOUS SYSTEM

Since the last century it has been usual to think of the nerves
as agents of communication, following the analogy of telegraph
wires. But what do they communicate? Neurophysiologists have
been cautious and confused about this ever since the time Des-

visual learning

-"^■C vertical lobe

touch learning

V'v--\'

(/ movement
} - of eyes
^and body •>

FIG. 2. Longitudinal sagittal section of the brain of an octopus.

cartes spoke of nerves with the analogy of pulling on wires to
ring bells or of animal spirits traveling along hollow tubes.

During the last century and the present one, physiologists
have mostly described the activity of nerves by using unques-
tioningly the phrases "nerve impulse" or "action potential," but
now we can see that these are rather ambiguous and indeed eva-

sive terms. This will sound like rank heresy, especially coming
from me since the giant nerve fibers of the squid have told us
more about nerve impulses than any other nerve fibers have
done (fig. 3). I came upon them by chance while studying squid
ganglia for another purpose. The cells related to them had in-
deed been seen by Williams in 1909. But there had been no fur-
ther mention of the cells in the literature, and no one had seen
the giant fibers themselves. In 1936 at Woods Hole we were
able to prove that these huge channels are nerve fibers and fig-
ure 4 shows some of the earliest records of their action poten-
tials. The function of these enormous nerves is to elicit contrac-
tion of the sac that produces the propulsive jet. The arrange-
ment ingeniously provides that both sides of the mantle and its
nearer and distant parts all contract together (fig. 5).

If the function is so well understood, what do I mean by say-
ing that the concept of an impulse or action potential is ambig-
uous? What's in a name? In this context of the giant fiber sys-
tem I agree that it does not matter much. An activity spreads

FIG. 3. Transverse section of one of the stellar nerves of a squid. There are many
small nerve fibers and one giant fiber.

FIG. 4. Oscillograph record of the electrical changes accompanying a sequence of
nerve impulses in a squid's giant nerve fiber. The discharge has been set off by plac-
ing oxalic acid on one end of the fiber. Note that the impulses are all the same height.
The time-markers show 1/5 or l/100th sec.

along the nerve fibers and we can tell rather precisely how it is
initiated by a synapse in the stellate ganglion and propagated to
start off a muscular contraction. We can even show that one
nerve impulse produces one pulse of the jet, so we can say that
the action of single cells in the nervous system produces a par-
ticular behavioral act by the whole squid. This is good progress
in understanding. We can go further and apply it to mammals
where, in a monkey trained to press a lever, single cells of the
cerebral cortex show electrical activity before the movement be-
gins (Evarts et al., 1971). Thus we get a good idea of how the
nervous system is made up of nerve cells each of which has a
distinct function.

This sounds fine and is indeed true. The principle on which
all nervous systems are built is that of multichannel communica-
tion. Each nerve fiber carries only one sort of message, either in-
ward from a sense organ or outward to produce some action by
a muscle or gland. Each fiber thus carries only a small amount
of information. To carry large amounts of information inward
and to produce varied and subtle behavior, very large numbers
of fibers are needed, each having a different "function." The
trouble is that in the more interesting parts of the brain we
cannot specify what the "function" is. So when we say that
when we see red certain nerve fibers from the eye transmit some-
thing called nerve impulses we do not really know what we
are saying. In what sense do nerve impulses transmit redness?

■a <n

To answer this we must look more carefully at what we mean
when we say that the nervous system serves for communica-
tion. We are using words borrowed from human activities in
which a sender has a message calling for some action that he
expects from a recipient. He passes signals in what we call a
code along a channel to the receiver, who decodes it and selects
the required action from the repertoire, or set, of programs
available. There are very many fascinating things we could say
about this situation. For the present, notice firstly that the ac-
tivity of communication presupposes an aim or purpose that is
to be achieved by choosing the right program from a set. Fur-
ther it makes use of some arbitrary code of signals, preset by
past history and "understood" by transmitter and receiver. Liv-
ing things are the only systems that we know of that maintain
themselves by communication in this way. So what we are
doing is to use the words that have been developed to describe
human social life to describe all living things. For the present
we are concentrating on nervous messages themselves, and we
notice that the analogy suggests that they be called signals in
a code. Physiologists are beginning to talk about nerve im-
pulses in this way but curiously enough the physiologists who
win Nobel Prizes for the study of nerve fibers seldom, or never,
use words such as "code" or "symbol." They stick to the dear
old terms "nerve impulse" and "action potential." They have in-
deed been able to find out a very great deal about the physical
changes that are involved in the transmission of the nerve mes-
sage, without thinking much about what the message communi-
cates. To be unkind one might say it was like giving a Nobel
Prize for Literature to people who had advanced knowledge of
typewriters, or of ink, or perhaps of radio transmission! I may
say that many of my best friends are Nobel Prize winners— at
least they have been until now!

But there are two further turns of the screw that physiolo-
gists must suffer. The significance of signals in a code is that
they symbolize the matters to be communicated. If we are to
describe the effects of our nerve impulses properly, in this
analogy we must say that they are significant because they are

symbols, that is, they stand for or represent either some event
in the outside world or some inner need or some action to be
performed at the decoding end of a communication channel.
We say that a sign or a signal becomes a symbol or representa-
tion for something else when it has the effect upon us of that
something. A traditional picture of a horse symbolizes horse for
us; but the horse in Picasso's "Guernica" does more, it sym-
bolizes also fear and horror.

I claim, therefore, that we shall learn to understand better
how the nervous system works if we consider how the opera-
tions of each part of it represent or symbolize either some
change in the inner or outside world or some instruction for
action, passing outward from the brain to the muscles or glands.
Let us then see what the various parts of the nervous system
in our cephalopods serve to symbolize.

SYMBOLS FOR GRAVILY AND MOVEMENT

Cephalopods, like other animals, arrange their behavior
in such a way as to respect the demands of gravity. To be able
to do this they have within themselves parts which by their
physical structure symbolize gravity and movement. These
are, as it were, little models of those features of the universe.
Cephalopod statocysts are based on principles surprisingly simi-
lar to those used by vertebrates, including man (fig. 6). Like our
own inner ear, they combine receptors for maintaining orienta-
tion with respect to gravity with others that are sensitive to the
angular accelerations due to movement of the animals.

The gravity receptors illustrate well the principles involved in
symbolization. To meet the task of correct orientation in rela-
tion to the earth's surface, there is present in the statocyst a
little model to represent gravity, a stone hanging upon sensory
hairs. These hairs send streams of action potentials whose pat-
tern thus symbolizes the position of the animal in relation to
gravity. The connections of these nerve fibers must be meticu-
lously arranged to ensure that the various muscles pull to pre-

mac. prin. mac. n. sup

FIG. 6. The statocyst of the squid Loligo, as seen from in front. The calcium salts
in the statoliths (the gravity stones) make them opaque. There is one very large one
on each side, composed of crystals of aragonite. This stone lies in the transverse plane
attached to sensory hairs of the macula princeps (mac. prin.). There are two other
patches of sensory cells, carrying numerous small crystals. The macula neglecta su-
perior (mac. n. sup.) lies nearly in the sagittal plane, the macula neglecta inferior
(mac. n. inf.) in an oblique horizontal plane.

cisely the correct extent to hold the animal upright (fig. 7). If
the statocysts are destroyed this is no longer possible. Notice,
then, that the model serves to allow the action system of the
animal to maintain its proper relation with the rest of the world—
the essential feature of living.

For the detection of angular accelerations the cephalopods
have ridges of sensory hairs, the cristae, carrying very light
flaps, the cupulae. The cristae run along the sides of the stato-
cyst sac in four directions, at right angles to each other (fig. 8).
When the animal turns, the displacement of the wall relative to
the fluid contents of the statocyst moves the cupula of one or
more of the ridges according to the direction of movement. The
signals set up by the hair cells of the crista thus represent the

FIG. 7. Drawings by M.T. Wells to show how pupil of a normal octopus is always
held horizontal (a-e). In f and g, are shown the positions of the pupils in an animal
from which both statocysts had been removed.

animal's own movements. By their connections these nerve fi-
bers then initiate compensatory movements, especially of the
eye muscles.

This system is obviously similar to that of our own semi-
circular canals. It is indeed striking that in the more active ceph-
alopods, such as the squids, the statocyst has become divided
up and curved into shapes that in effect constitute actual
canals. Our three semicircular canals serve to represent angular
accelerations in three planes of space. What are the squids doing
with four cristae? It may be that the answer is that with the
fourth they detect linear acceleration forward or backward.
These animals can move readily in these two directions, which
is a feat not easily achieved even by their rivals the fishes.
Budelmann (1975) has shown that the cristae are indeed capa-
ble of responding to linear acceleration (unlike the semicircular
canals).

a.c.

FIG. 8. Statocyst of the fast -moving squid Loligo, seen from above. The stato-
liths shown in figure 6 have been removed. The white outlines show the course of the
crista (ridge) of sensory cells (cr) mainly for detecting angular accelerations. The
ridge runs (on each side) across in front, along the side, across the back and then
up in the vertical plane. The cavity has curved sides and is divided up by a number of
projections (a.c.=anticristae). The effect is a restriction of fluid movement similar to
that accomplished by the semicircular canals in vertebrates.

It is interesting to note that in octopuses and other cephalo-
pods that do not make rapid turning movements the whole sys-
tem is changed. The sac is very large and the anticristae are re-
duced or absent, leaving a single volume of fluid whose inertia
gives greater sensitivity to slow movements (fig. 9). So in every
animal the structure and connections of the sense organs have
come to represent the environment in which it lives. Notice that
the model that the animal contains represents not only the fea-
tures of the world but also the actions that the animal must it-
self perform to keep alive. The models in the brain are not static
pictures, they are the written plans and programs for action. In
squids the giant cells that produce the jet lie very close indeed
to the statocyst. If the animal is suddenly disturbed it imme-
diately produces a jet. This plan of action does not have to

FIG. 9. The statocyst of the slow-moving squid Taonius, seen from above. The
macula (mac.) and its stones are quite different from those of Loligo. The sac is
large and the anticristae (a.c.) are small and few, so that the cavity is not divided up
into "canals." K. is Kolliker's canal, a blind ciliated tube of unknown function ;cr =
crista.

be learned. It is written into the inherited wiring pattern.

In man and other vertebrates the cerebellum is a very im-
portant part of the system for control of movement. We have
recently realized that there are lobes in the brains of cephalo-
pods that contain large numbers of very small parallel fibers,
strikingly like those of our own cerebellum (figs. 10, 11). We
do not yet understand the full significance of these arrange-
ments but a possible explanation is that the fine fibers serve
to represent time (Braitenberg, 1967). They conduct very
slowly and this may determine the braking action that termi-
nates a movement. Many actions of the muscles are ballistic,
in the sense that the ending of their contraction is determined
when it begins and not by any feedback en route.

In ourselves the ear has the further function of detecting
sound. Cephalopods seem to have no capacity for responding
to vibrations, except those of very low frequency. This is very
strange since water transmits vibrations that could have very

med. bas
fins
chromatophores

mantle

eye-muscles

funne

FIG. 10. A diagram of the brain of a squid showing the four sets of fine parallel
fibers, somewhat similar to those in the vertebrate cerebellum. The suboesophageal
lobes lie below and control the various movements as shown. The cerebellum-like
lobes lie above them and are called the anterior basal (a.bas.), median basal (med.
bas.) and peduncle lobes (ped.). The parallel fibers run in different planes; two sets
are in the anterior basal lobe, one in each of the others. Notice that these lobes send
fibers to the lower motor centers.

great symbolic value and indeed the fishes, great rivals of the
cephalopods for domination of the waters, hear very well.

LEARNING SYMBOLIC VALUES

All the behavioral responses we have considered so far have
been the consequence of connections laid down during develop-
ment, but cephalopods are provided also with considerable
powers of learning. Far less of course than in mammals or man
but still enough to provide us with much information about the
processes that are involved in memory formation. It is here
that it becomes especially important to pay attention to our
conceptual framework and language. The essence of learning is
the attaching of symbolic value to signs from the outside world.
Images on the retina are not eatable or dangerous. What the eye

FIG. 1 1. Section of the peduncle lobe of a squid showing the fine parallel fibers.
Stained by the Golgi method, which picks out a few fibers. The photograph has been
retouched.

can provide is a tool by which, aided by a memory, the animal
can learn the symbolic significance of events. The record of its
past experiences then constitutes a program of behavior appro-
priate for the future.

Octopuses have two separate memory systems. One allows
them to make appropriate responses to things that they see; the
other does the same for the tactile and chemical properties of
objects touched by the arms (fig. 12). These systems lie at the
top of the hierarchy of nerve centers in the sense that they
make the decisions as to which movements shall be executed by
the lower parts. To revert to our military metaphor, they are
the General Staff. They receive intelligence from the outside
world and then write plans for programs of action by the whole
army, in the light of their memory records of past experience.

With the visual system an octopus can learn to make attacks
at one shape but to retreat from another. With the touch system

vert. sup. fr

Attack" /"| ^V~?

raw in" inf. fr.

optic lobe

eye

arms

funnel

FIG. 12. Diagram of the brain of an octopus showing the parts that make up the
two memory systems. The two are outgrowths from the superior buccal lobe, which
controls the eating system (sup. bucc). The inferior frontal system (inf. fr.) receives
information from the arms and provides a memory regulating which objects are
drawn in. The superior frontal (sup. fr.) and vertical (vert.) lobes are part of the visual
memory, serving to decide which objects should be attacked for food.

he can learn to discriminate degrees of roughness and also
chemical differences, detected by the suckers (Wells and Wells,
1956; Wells, 1963) (fig. 13).

The visual system has features again surprisingly like those of
vertebrates in their principles of operation, in spite of great dif-
ferences in detailed anatomy. We can see from these principles
the stages that are necessary for the learning of symbolic signif-
icances by vision or touch.

FEATURE DETECTORS

The first essential is to have sensors that are competent to
extract relevant information from the world. We know little
about the physiology of these in cephalopods but something of
their anatomy. There are cells with receptive fields in the outer
parts of the optic lobes that seem suited to detect contours, as

<^<*NrH £*}£*!

i^|% g?§f*%

FIG. 13. Series of plastic spheres used for training octopuses to distinguish various
degrees of roughness.

do cells of the visual cortex of mammals (fig. 14). Octopuses
can be trained to react differentially to rectangles with vertical
and horizontal orientations. It is probable that these features
are detected by the receptive fields of these second-order visual
cells, which seem to be tuned to receive signals from rows of
optic nerve fibers. We note that such a system depends on a de-
tailed somatotopic projection from the sensory surface of the
eye. This presents a literal map of outside events, from which
the brain then records certain features as it writes the programs
that will determine its future actions. Moreover, these feature
detectors lie in a layered system of neuronal processes, the
plexiform layer, which is surprisingly like the layered structure
of the vertebrate retina (fig. 15). Contributing to this layered
neuropil are great numbers of amacrine and horizontal cells,
with processes limited to the plexiform layer. Some extend over
long distances, others are quite short, and we have as yet no
information as to how any of them operate. Their presence,
however, in essentially the same relations in cephalopods and

Res-

Res.+

FIG. 14. Diagram of the optic lobe of an octopus to show the system by which it
is suggested that visual contours are detected and memory records made that will
control future behavior. Clas. V. and clas. H. are the "classifying cells," which
respond to particular visual features (e.g., vertical or horizontal rectangles). The
octopus can be trained to attack or avoid either of these, so the pathways from them
msut lead to motor systems for attack and retreat. Following an attack the animal
will receive either food or pain. The suggestion is that signals from the lips (food) or
from the body (pain), besides promoting attack or retreat, will activate the small
cells, which produce an inhibitory transmitter and block the unwanted pathway,
leading to greater use of that which is "correct." The memory cells (mem.) only
discharge if they receive signals both from the classifying cells and from the indicators
of results (Res.+ and Res.-). The system is shown biased as it would be if the horizon-
tal rectangle had been given food and the vertical shocks.

vertebrates should surely help us to find the principles that are
involved in the extraction of significant visual features. Pribram
(1971) has suggested that such systems recall the logical organi-
zations necessary for encoding by and/or gates. We can also sur-
mise from the work of Dowling and Werblin (1969) on the
retina of the mud-puppy (Necturus) that these elaborate net-
works operate essentially as analogue computers, using patterns
of graded electrical signals to compute from the patterns that
are sent to them from the retinal receptors suitable all-or-none
signals to pass on to the next stage in the brain.

Unfortunately, we know rather little about how to pursue
such signals, either in cephalopods or vertebrates, to the points

FIG. 15. Photograph of a section of the surface of the optic lobe of an octopus,
showing how it resembles the vertebrate retina. There are outer and inner granule cell
layers (o. gr. and in. gr.), with a plexiform layer between (plex.). The optic nerve
fibers come in from the right (o.n.). They have disappeared from the upper part of
the figure where some of them had been cut some days previously. The inner
tangential bands of fibers in the plexiform zone (tan.) are the receiving dendrites of
the "classifying cells" shown in figure 14. They have remained intact. Cajal's silver
stain.

at which the changes occur that constitute the writing of a new
action program by the memory mechanism. In squids we can
say that there are only one or two further synapses between the
feature detectors and the giant cells. Therefore, although the
optic lobes are indeed large and complex, there is no need to
suppose that any very elaborate system of operations has to in-
tervene between detection and behavior, even in learned behav-
ior.

However, somewhere in this pathway there must be the possi-
bility of an alteration in connection patterns, if that is the
mechanism by which the memory system works. I have sug-
gested that this is done by the operation of a switch system that
reduces the probability of using one pathway in favor of the
other (fig. 14). It may be that once one path begins to be used
rather than the other there will also be a subsequent increase in
its availability, perhaps by added synaptic connections or effi-
cacy, as has been suggested, following Cajal (1895, see 1953 p.
887), by many workers (e.g., Hebb, 1949; Young, 1950). But
whatever mechanism is used to establish the symbolic value of
some set of nervous signals, it must involve a reduction of the
number of possible behavioral responses. The octopus can orig-
inally react either positively or negatively to a horizontal
rectangle; his experience restricts him to only one of these re-
sponses. A given signal cannot symbolize both something good
and bad. I have suggested that the switching of each single neu-
ronal pathway constitutes a unit of memory or mnemon. It is
the single "word" of the writing that constitutes the new pro-
gram of action. The octopus is a very simple creature and per-
haps it learns only single words. We have to learn not only
words but whole "sentences," indeed whole "books," which
constitute the action programs that become written in our
memories.

For the establishment of symbolic value it is essential that
the results of action can be referred to a standard, which must
ultimately be set by the genetic composition, the historical in-
formation encoded in the DNA. Such signals of the results of
action come from the taste systems on the one hand and the

pain systems, producing aversive responses, on the other. We do
not know much about them in octopuses but there is evidence
that if they are prevented from reaching to the appropriate
parts of the brain no learning is possible. We notice that these
nerve impulses, like all others, are symbolic, in this case sym-
bolizing internal states that are either satisfactory or unsatis-
factory for life. The symbolic value is established by the long
sequence of selections that have produced appropriate DNA.
Those organisms that do not have an appropriate taste for food
and life or skill in avoiding pain do not survive.

The anatomy suggests that in the octopus, as in vertebrates,
special patterns of connection are used to allow these reference
signals to meet with those coming from the outside world. In
both the visual and touch memory systems of the octopus there
are lobes in which this interaction can take place (fig. 16). The

sup.fr. -^

-.1 .

*%i&

FIG. 16. Photograph of sagittal section through the front part of the brain of an
octopus, showing the inferior frontal (inf. fr.), superior frontal (sup. fr.) lobes, and
superior buccal lobe (sup. bucc). These serve to mix signals of taste (from the lips)
with those from the arms and optic lobes (respectively). The two lobes have similar
structures, with many interweaving bundles, allowing for the mixing. Cajal's silver
stain.

output of the lobes in both cases passes through a further lobe
consisting of large numbers of very small cells, the vertical or
subfrontal lobes (fig. 17). Many lines of investigation have
shown that these lobes are involved in the process of recording
in the memory, but are not absolutely essential for it. Their
action seems to be particularly in restraining the animals from
performing actions that are likely to be damaging. The numer-
ous minute cells in these lobes can be seen with the electron
microscope to be packed with synaptic vesicles (fig. 18). How
they operate remains a very interesting question.

In general we can say that if learning consists in increasing
the probability of performing certain "correct" actions when
symbols appear, then it is necessary to have inhibitory systems
to restrain the performance of other actions. A multichannel
system such as this operates by means of a maximum ampli-
tude filter in which many elements may be active but only the
most active takes control (Taylor, 1964). It is suggested that the
cerebral cortex contains systems that act in this way. Perhaps
the prefrontal lobes in particular have a restraining influence in

pain

FIG. 17. Diagram of some connections of the median superior frontal (med. sup.
fr.) and vertical lobes (vert.) of an octopus as shown by electronmicroscopy. The
short amacrine cells in the vertical lobes are packed with synaptic vesicles. They are
influenced by the fibers from the superior frontal and also by those entering from
below and probably signalling pain. They influence larger cells leading to the
subvertical lobe (subv.) and so back to the optic lobes (opt.).

WT&

FIG. 18. Electronmicrographs of synaptic contents in the superior frontal (on
right) and vertical (left) lobes of an octopus. The synapses in the former are between
incoming fibers (e) and cell processes (cp.). In the vertical lobe the amacrine trunks
(amt.) receive synapses from the axons of the superior frontal (s.f.b.) and transmit to
spines (sp.) of cells that carry signals away from the lobe.

man, allowing the performance of such delicately graded actions
as those of effective speech in a social context.

Human brains, like those of octopuses, must contain reference
systems to determine which lines of action are likely to be suc-
cessful in maintaining life. We can indeed begin to see some evi-
dence that they operate in ways rather like those described.
Ungerstedt (1971) and others have shown that there are systems
of aminergic pathways leading upward from centers in the
medulla to the hypothalamus and on to the limbic system and
frontal cortex (fig. 19). These pathways, such as that beginning
in the nucleus coeruleus, come from regions where fibers from
the taste buds enter the brain. Crow and his colleagues have pro-
duced evidence that rats with lesions to this pathway cannot
learn to run a maze for food reward (Anlezark, Crow, and
Greenaway, 1973). Moreover, with electrodes implanted in
these regions animals will press repeatedly for self-stimulation.
There are controversies about these experiments, but it seems
very probable that we are approaching here close to the core of
many problems that have worried mankind for centuries, and
do so still. The reference signals that come from these path-
ways, and from the hypothalamus, provide the aims and objec-
tives of our lives and the course of our learning. Of course crude

loc. c.

FIG. 19. Diagram of the ascending pathways on the rat's brain that use the
transmitter noradrenaline. They begin in the locus coeruleus (loc. c.) and other
centers in the hind brain. From here they ascend to the cerebellum (cb.), hypo-
thalamus (hyp.) and finally reach to the cerebral cortex (cort.), olfactory bulb (ol.)
and hippocampus (hip.). The terminal areas are shaded (after Ungerstedt, 1971).

rewards do not necessarily enter into every associational act,
especially in man. We have acquired more subtle systems of re-
ward to supplement those of taste and pain. Nevertheless, we
begin to see how life depends upon symbolic signs of life values,
which are used to give symbolic significance to the signals we
receive from the outside world.

LITERATURE CITED

Anlezark, G.M., T.J. Crow, and A. P. Greenaway

1973. Evidence that noradrenergic innervation of the cerebral cortex is necessary
for learning. Jour. Physiol. London, vol. 231, pp. 119-120.

Arbib, M.A.
1972. The metaphysical brain. New York, WileyTnterscience.

Boycott, B.B.

1961. The functional organization of the brain of the cuttlefish, Sepia officinalis.
Proc. Royal Soc, vol. 152, pp. 503-534.

Boycott, B.B. and J.Z. Young

1950. The comparative study of learning. Symp. Soc. Exp. Biol., vol. 4, pp. 432-
453.

Braitenberg, V.

1967. Is the cerebellar cortex a biological clock in the millisecond range? Progr.
Brain Res., vol. 25, pp. 334-346.

Budelmann, B.-U.

1975. Gravity receptor function in cephalopods with particular reference to
Sepia officinalis. Fortsch. Zool., vol. 23, pp. 84-96.

Cajal, S.R.
1953. Histologie du Systeme nerveux (facsimile of French edition of 1909). Ma-
drid, Instituto Ramon y Cajal.

Dovvling, J.E., and F.S. Werblin

1969. Organization of retina of the mudpuppy Necturus nee turns. 1. Synaptic
structure. Jour. Neurophysiol., vol. 32, pp. 315-338.

Evarts, E.V., E. Bizzi, R.E. Burke, M. Long, and W.T. Thach, Jr.

1971. Central control of movement. Neurosci. Res. Program Bull., vol. 9, pp. 2-
170.

Hebb, D.O.

1949. The organization of behavior. New York, John Wiley.
Pribram, K.H.

1971. Languages of the brain. New Jersey, Prentice Hall.
Taylor, W.K.

1964. Cortico-thalamic organization and memory. Proc. Royal Soc. B., vol. 159,
pp. 466-478.

Ungerstedt, U.

1971. Stereotaxic mapping of the monomine pathways in the rat brain. Acta
Physiol. Scand., vol. 367 (Suppl.), pp. 148.

Uclls. M.J.

1963. Taste or touch: some experiments with Octopus. Jour. Exp. Biol., vol. 40,
pp. 187-193.

Wells. M J. .and J.Wells

1956. Tactile discrimination and the behavior of blind Octopus. Pubbl. Staz.
zool. Napoli., vol. 28, pp. 94-126.

Williams, L.W.

1909. The anatomy of the common Squid Loligo pealii, Lesueur. Leiden, Brill.

Young, J.Z.

1939. Fused neurons and synaptic contacts in the giant nerve fibres of cephalo-
pods. Phil. Trans. Royal Soc. vol. 229, pp. 465-503.

1950. Doubt and certainty in science. Clarendon Press, Oxford.

1971. The anatomy of the nervous system of Octopus vulgaris. Oxford, Clarendon
Press.

SN2&1.4
• J35
10.4-7
1977

FORTY-SEVENTH
I JAMES ARTHUR LECTURE ON
THE EVOLUTION OF THE HUMAN BRAIN

AN EVOLUTIONARY INTERPRETATION

OF THE
PHENOMENON OF NEUROSECRETION

BERTA SCHARRER

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1977

*fi

1869
THE LIBRARY

FORTY-SEVENTH

J WHS ARTHUR U C I I Rl ON

THE EVOLUTION OF THE HUMAN BRAIN

AN EVOLUTIONARY INTERPRETATION

OF THE
PHENOMENON OF NEUROSECRETION

BERTA SCHARRER

Professor of Anatomy and Seuroscience
Albert Einstein College of Medicine. New York

THE AMERICAN MUSEUM OF NATURAL HISTORY
NEW YORK : 1978

FORTY-SEVENTH

JAMES ARTHUR LECTURE ON

THE EVOLUTION OF THE HUMAN BRAIN

The author's studies reported in the present paper have been
supported by Research Grants NB-05219, NB-00840, and 5 POl-
NS-07512 from the U.S.P.H.S. and by N.S.F. Grant BMS
74-12456.

The treatment of the literature was greatly facilitated by Bibli-
ographia Neuroendocrinological compiled and edited by Dr.
Mary Weitzman, and supported by Grant No. 1 ROl LM 02327
awarded by the National Library of Medicine, U.S.P.H.S.,
DHEW.

JAMES ARTHUR LECTURES ON
THE EVOLUTION OF THE HUMAN BRAIN

Frederick Tilney. The Bruin in Relation to Behavior, March 15, 1932

C. JtldsOi] Herrick, Brains as Instruments of Biological Values; April 6. 1933

D. M. S. Watson. The Story of Fossil Brains from Fish to Man; April 24, 1934

C. U. Aliens Kappers, Structural Principles in the Nervous System; The Development of
the Forehrain in Animals and Prehistorie Human Races; April 25, 1935

Samuel T. Orton, The Language Area of the Human Brain and Some of its Disorders;
May 15. 1936

R. W. Gerard, Dynamic Neural Patterns; April 15, 1937

Franz Weidenreich. The Phylogenetic Development of the Hominid Brain and its Connec-
tion with the Transformation of the Skull; May 5, 1938

G. Kingsley Noble, The Neural Basis of Social Behavior of Vertebrates; May 11, 1939

John F. Fulton, A Functional Approach to the Evolution of the Primate Brain; May 2,
1940

Frank A. Beach, Central Nervous Mechanisms Involved in the Reproductive Behavior of
Vertebrates; May 8, 1941

George Pinkley. A History of the Human Brain; May 14, 1942

James W. Papez, Ancient Landmarks of the Human Brain and Their Origin; May 27,
1943

James Howard McGregor, The Brain of Primates; May 11, 1944

K. S. Lashley, Neural Correlates of Intellect; April 30, 1945

Warren S. McCulloch, Finality and Form in Nervous Activity; May 2, 1946

S. R. Detwiler, Structure-Function Correlations in the Developing Nervous System as
Studied by Experimental Methods; May 8, 1947

Tilly Edinger, The Evolution of the Brain; May 20, 1948

Donald O. Hebb. Evolution of Thought and Emotion; April 20, 1949

Ward Campbell Halstead, Brain and Intelligence; April 26, 1950

Harry F. Harlow. The Brain and Learned Behavior; May 10, 1951

Clinton N. Woolsey, Sensory and Motor Systems of the Cerebral Cortex; May 7, 1952

Alfred S. Romer, Brain Evolution in the Light of Vertebrate History; May 21, 1953

Horace W. Magoun, Regulatory Functions of the Brain Stem; May 5, 1954

*Fred A. Mettler, Culture and the Structural Evolution of the Neural System; April 21,
1955

*Pinckney J. Harman, Paleoneurologic, Neoneurologic, and Ontogenetic Aspects of Brain
Phylogeny; April 26, 1956

*Davenport Hooker, Evidence of Prenatal Function of the Central Nervous System in
Man; April 25, 1957

*David P. C. Lloyd, The Discrete and the Diffuse in Nervous Action; May 8, 1958

*Charles R. Noback, The Heritage of the Human Brain; May 6, 1959

*Ernst Scharrer, Brain Function and the Evolution of Cerebral Vascularization; May 26,
1960

Paul I. Yakovlev, Brain, Body and Behavior. Stereodynamic Organization of the Brain
and of the Motility-Experience in Man Envisaged as a Biological Action System; May
16, 1961

H. K. Hartline, Principles of Neural Interaction in the Retina; May 29, 1962

Harry Grundfest, Specialization and Evolution of Bioelectric Activity; May 28, 1963

*Roger W. Sperry, Problems Outstanding in the Evolution of Brain Function; June 3,
1964

*Jose M. R. Delgado, Evolution of Physical Control of the Brain; May 6, 1965

Seymour S. Kety, Adaptive Functions and the Biochemistry of the Brain; May 19, 1966

Dominick P. Purpura, Ontogenesis of Neuronal Organizations in the Mammalian Brain;
May 25, 1967

*Kenneth D. Roeder, Three Views of the Nervous System; April 2, 1968

tPhillip V. Tobias, Some Aspects of the Fossil Evidence on the Evolution of the Hominid
Brain; April 2, 1969

*Karl H. Pribram, What Makes Man Human; April 23, 1970

Walle J. H. Nauta, A New View of the Evolution of the Cerebral Cortex of Mammals;
May 5, 1971

David H. Hubel, Organization of the Monkey Visual Cortex; May 11, 1972

Janos Szentagothai, The World of Nerve Nets; January 16, 1973

*Ralph L. Holloway, The Role of Human Social Behavior in the Evolution of the Brain;
May 1, 1973

*Elliot S. Valenstein, Persistent Problems in the Physical Control of the Brain; May 16,
1974

a Marcel Kinsboume, Development and Evolution of the Neural Basis of Language; April
10, 1975

*John Z. Young, What Squids and Octopuses Tell Us About Brains and Memories, May
13, 1976

*Berta Scharrer, An Evolutionary Interpretation of the Phenomenon of Neurosecretion;
April 12, 1977

^Unpublished.

*Published versions of these lectures can be obtained from The American Museum of
Natural History, Central Park West at 79th St., New York, N. Y. 10024.

tPublished version: The Brain in Hominid Evolution, New York: Columbia University
Press, 1971.

AN EVOLUTIONARY

INTERPRETATION OF THE

PHENOMENON OF NEUROSECRETION

INTRODUCTION

Almost 50 years ago, Ernst Scharrer (1928) made a discovery
that was received by the scientific community with great skepti-
cism, if not with outright rejection. It marked the beginning of a
scientific adventure that has given rise to one of the most chal-
lenging pursuits in neurobiological research, the results of which
have been dramatic. Based on cytological observations in a tele-
ost fish, Phoxinus laevis, he postulated that certain groups of
distinctive cells in the hypothalamus ("neurosecretory neurons")
engage in secretory activity to a degree comparable to that of
endocrine gland cells. He further suggested that this activity may
be related to hypophysial function.

A search in the literature yielded but one comparable report
calling attention to the occurrence of "glandlike" nerve cells in
another part of the central nervous system, the spinal cord of
skates (Speidel, 1919). Subsequent studies demonstrated the al-
most ubiquitous occurrence of such neurons throughout the ani-
mal kingdom. Yet, for many years to come, the spotlight
remained on the hypothalamic neurosecretory centers of the ver-
tebrate series. The elucidation of their close affiliation with the
pituitary gland eventually gave rise to a new discipline, neuroen-
docrinology.

However, recognition of these unusual neural elements repre-
sents a challenge to the neuron doctrine, according to which
nerve cells are most commonly thought of as being designed for
the reception of stimuli, the generation and propagation of bio-
electrical potentials, and the rapid, synaptic transmission of sig-
nals to contiguous recipient cells. "Conventional" neurons make
only very restricted use of chemical mediators in the form of
special neurotransmitters and certain other regulatory substances.

In contrast, the primary activity of the classical neurosecretory
cell consists of the manufacture of a distinctive (proteinaceous)
product in sufficient quantity to function in a hormonal capacity.
Furthermore, its axon terminates, without establishing synaptic
contact, in close proximity to the vascular system. No wonder
that, because of these "aberrant" attributes, neurosecretory phe-
nomena were long looked at askance and frequently brushed off
as signs of degenerative or postmortem changes. Today, this
mistrust no longer exists, as will become apparent in the follow-
ing discussion.

An important step forward was the demonstration that the
"posterior lobe hormones," e.g., vasopressin, are derived from
peptidergic neurosecretory nuclei of the hypothalamus and are
transported by axoplasmic flow to the posterior pituitary where
they are released into the general circulation (Bargmann and
Scharrer, 1951). At last, the endocrine nature of some of these
neuroglandular elements had been established.

But even then a vexing question remained. Why should the
body make use of nerve cells to provide hormonal messengers in
order to reach terminal effector sites such as the kidney? This
conceptual difficulty stems from the established custom of classi-
fying neuroculatory and glandular functions as two distinctly
separate categories. In reality, there are cogent reasons for bridg-
ing this gap, based largely on evidence that is implicit in the
evolutionary history of integrative systems.

PHYLOGENY OF NEURAL SYSTEMS OF COMMUNICATION

A careful consideration of the phylogenetic, and to some ex-
tent the ontogenetic, development of such informational systems
has made it increasingly clear that the manufacture and release of
secretory products is an old and fundamental attribute of neuronal
elements. The evolutionary approach, featured in this article, is
based on the premise that the most elementary integrative mecha-
nisms existing today may resemble those of our remote ancestors
(Pavans de Ceccatty, 1974).

By reason of its phylogenetic derivation from a pluripotential
epithelial element, the primitive nerve cell can be viewed as a

functionally versatile structure, endowed with the capacity to
dispatch both long distance and localized chemical signals. This
concept is supported by a substantial body of information on
neuroregulatory mechanisms encompassing all multicellular ani-
mals (see Lent/., 1968; Highnam and Hill, 1977). Starting with
the simplest forms among them, we find in sponges a reticular
neuroid tissue complex whose components do not yet satisfy all
the criteria of nerve cells. The first primitive neurons with ele-
mentary synaptic contacts appear in the lowest eumetazoans, the
eoelenterates.

What seems important in the context of the present analysis is
that in both groups some of the cells mentioned display cytologi-
cal signs of neurosecretory activity. The cytoplasmic granules
observed here are comparable to those of higher animals in that
they stain with alcian blue and, in electron micrographs, appear
electron dense and membrane-bounded with diameters of
1,000-1.700 A (Pavans de Ceccatty, 1966; Lentz, 1968; Davis,
1974). Neurosecretory granules are abundant in the nervous sys-
tem of planarians. In the ganglia of annelids more than one-half
of all neurons are of the neurosecretory type.

Even more relevant is the fact that distinctive hormonal func-
tions as well as other "nonconventional" neuroregulatory roles
can be ascribed to the neurosecretory neurons of these lower
invertebrates. Tests with isolated neurosecretory granules of the
coelenterate Hydra reveal that their content regulates growth and
differentiation, especially during regeneration (Lentz, 1968). This
neuromediator also seems to participate in the induction of
gametogenesis and sexual differentiation (Burnett and Diehl,
1964). Similarly, neurosecretory control, over a distance, of
growth during development and regeneration and of certain re-
productive events has been demonstrated in planarians (Lentz,
1968; Grasso and Benazzi, 1973) and in annelids (Hauenschild,
1974).

The salient point is that in none of these primitive invertebrates
have "regular," i.e., nonneural glands of internal secretion been
identified. Therefore, at this level of differentiation, the nervous
system seems to be the only agency available for carrying out all
of the existing endocrine functions. Neurohormones thus hold the

rank of the phylogenetically oldest integrative long-distance mes-
sengers, and the endocrine type of coordination accounts for a
relatively large sector of neuronal activities in lower inverte-
brates. In other words, far from being a latecomer and a rare
exception, the neurosecretory neuron dates back to the very be-
ginning of the development of neural structures. Furthermore, its
versatility indicates that it has remained closer to the nerve cell
precursor than has the more specialized "conventional" neuron.

In the course of evolution the scene shifts in more than one
direction. Not only is there a staggering increase in the number
of neurons, as primitive nervous systems give way to more and
more elaborate structures but, in the most advanced forms, the
vast majority of "conventional" nerve cells engage in interneuro-
nal synaptic transmission involving the release of tiny, precisely
metered amounts of chemical transmitter substances. Some of
these special messengers are used over and over again. There-
fore, in these billions of neurons, the demands for secretory
activity have become greatly reduced.

Another, equally important, evolutionary change in design is
the evolvement of extraneuronal hormone sources. The structural
and functional attributes of the endocrine apparatus have long
been thought to be as clearly defined as have those of neurons.
Yet, the dividing line is by no means complete, on account of the
neuroectodermal origin of peptide-producing endocrine cells to be
discussed below. According to this concept (Pearse, 1976), the
relationship between the neuronal and nonneuronal hormone
sources of the hypothalamic-hypophysial complex is even closer
than formerly recognized, since both share their origin from the
same precursor cells in the ventral neural ridge. However, in
spite of this embryonic background, the cells of the ade-
nohypophysis should not be classified as neural elements. They
have crossed over to join the ranks of the endocrine system. This
process of metamorphosis entails the loss of structural and
cytochemical neural attributes and the acquisition of endocrine
qualities which these polypeptide-hormone-secreting cells share
with the rest of the endocrine apparatus.

The evolvement of this second integrative system that special-
izes in hormonal communication, making use of various types of

chemical messengers, argues against the need for blood-borne

neurochemical mediators in higher animals. Quite obviously, its
existence should relieve neurons from doing double duty.

NEUROENDOCRINE INTERACTIONS

In reality neurohormones do not become obsolete after the
acquisition, by arthropods and vertebrates, of an endocrine appa-
ratus proper. Instead they take over a novel and highly significant
role, that of mediation between the two systems of integration.
As has been pointed out repeatedly in the past (Scharrer,
1970-1974), the neurosecretory neuron, having retained its dual
capacity, is ideally suited and programmed for this special task.

In view of this shift in functional significance, the question
raised earlier, concerning the raison d'etre of first-order neurohor-
monal mechanisms even in the most highly developed organisms,
now appears in a different light. Such one-step systems, e.g., the
control of water metabolism by vasopressin, have certainly be-
come overshadowed by those constituting the all-important neu-
roendocrine channel of communication. They do not even seem
to be obligatory. Yet, their existence makes sense in an evolu-
tionary perspective, i.e., when interpreted as carryovers from
systems operating by necessity in phylogenetically less advanced
forms.

NONNEUROHORMONAL PEPTIDERGIC ACTIVITIES

A rather unexpected and challenging result of the detailed
ultrastructural analysis of the neuroendocrine axis in vertebrates
and invertebrates was the realization that not all the neurosecre-
tory neurons dispatch their messenger substances via the general
or the special portal circulation. There are, in fact, several recog-
nized modes of neurochemical communication that are neither
strictly neurohumoral (synaptic) nor neurohormonal (blood-
borne). One such mechanism is the long known regulation of
tissue growth and maintenance, by "neurotrophic substances"
(see Smith and Kreutzberg, 1976). The chemistry and extracellu-
lar pathway of these diffusible substances released from sensory
and motor fibers are still uncertain.

Information on nonvascular extracellular avenues available to
peptidergic neurosecretory messengers is more precise. These
variants include the cerebrospinal fluid (see Rodriguez, 1976),
zones of extracellular stroma, and even narrow "synaptoid" gaps.
Axons laden with neurosecretory material can be observed to
penetrate the glandular parenchyma of the adenohypophysis as
well as the corpus allatum of insects. In both organs synaptoid
release sites occur in close vicinity to, or even in contiguity with,
their apparent cells of destination. Furthermore, such spatial rela-
tionships are not restricted to endocrine elements, but are also
found in a variety of somatic structures, among them various
exocrine gland cells and muscle fibers.

Perhaps the most unexpected informational systems are those
in which neurosecretory neurons establish synapse-like relation-
ships with other neurons, some of which may themselves be of
the nonconventional type (see Scharrer, 1976). The realization
that, at least in certain special situations, peptidergic neurosecre-
tory mediators may operate in a manner comparable to that of
neurotransmitters has added a new and important facet to the
"gestalt" of the classical neurosecretory neuron. The existence of
these several intermediary possibilities for the transfer of informa-
tion by neurosecretory cells has clarified their relationship with
the more conventional neuronal types. Consequently, the sharp
dividing line originally thought to separate conventional from
classical neurosecretory neurons no longer exists. Now the modes
of operation of classical neurosecretory neurons actually blend
into a continuum of diverse neurochemical activities.

NONCONVENTIONAL INTERNEURONAL COMMUNICATION

What makes the discovery of synaptoid structures between
neurons intriguing is that they go hand in hand with increasing
physiological evidence in support of the concept that nonconven-
tional (peptidergic) neuroregulators may modulate certain forms
of synaptic interneuronal communication. This broader neu-
rotropic activity will undoubtedly turn out to represent a novel
and important form of information transfer with far-reaching bio-
medical implications (see, for example, Constantinidis et al.,

1974; Sterba. 1974; Broun and Vale. 1975; Plotnikoff ct al..
1975; Prange et al.. 1975a. 1975b; Vincent and Arnauld. 1975;
Brownstein et al.. 1976; Guillemin. Ling and Burgus. 1976; Lote
et al.. 1976).

To cite an example of such known activities among the hypo-
physiotropic hormones, or factors. TRF (thyrotropin releasing
factor) has a modulating effect on synaptic, especially mono-
aminergic. transmission. Apparently, this role evolved before that
of controlling thyrotropin release, and it seems to be of a more
general importance (Grimm-J0rgensen. McKelvy and Jackson.
1975; McCann and Moss. 1975; Waziri, 1975; see also Nicoll,
1977). A broader role for the posterior lobe hormone vasopressin,
or fragments thereof, is that demonstrated by de Wied and his co-
workers (1976) and involved in the control of various forms of
behavior. Moreover, effects that differ from conductance changes
evoked by conventional neurotransmitters can be elicited in cer-
tain neurosecretory neurons of molluscs by the application of
vasopressin and related peptides (Barker and Gainer. 1974). Fi-
nally, there is new and intriguing evidence that two specific
neuronal pentapeptides (enkephalins. Hughes et al.. 1975) func-
tion as endogenous analgesics presumably by suppressing excita-
tory synaptic signals implicated in the perception of pain (see
Snyder, 1977).

BIOCHEMICAL EVOLUTION OF NEUROSECRETORY
MEDIATORS

The evolutionary interpretation of the phenomenon of neu-
rosecretion presented here is not based on morphological and
physiological evidence alone. It can be further substantiated by
tracing the biochemical history of neurosecretory mediators, even
though the picture is still incomplete.

Among the general trends that are beginning to emerge are the
following. In contrast to those used in much smaller amounts by
conventional nerve cells, the chemical messengers operating in
classical neurosecretory neurons of both vertebrates and inverte-
brates are proteinaceous in nature. Furthermore, the biologically
active polypeptides of many neurosecretory neurons are bound by

noncovalent forces to special carrier proteins, called neurophysins
(see Walter, 1975; Watkins, 1975; Acher, 1976b), which are
primarily responsible for the selective stainability of neurosecre-
tory material throughout the animal kingdom. Aside from serving
as carrier molecules, these proteins may play an active role of
their own (Pilgrim, 1974).

Much information is being amassed on the occurrence and
precise localization of such neuropeptides and their affiliated
neurophysins within the neurosecretory systems of a variety of
animals by the use of immunochemical, especially immu-
noelectron-microscopic methods (see McNeill et al., 1976; Ude,
1976; Zimmerman, 1976). In addition, synthetically produced
neurohormones and their analogs are becoming available in in-
creasing numbers. These advances offer valuable tools for the
differential determination of the relationships and functional roles
of these substances.

There is substantial support for the concept that the characteris-
tic products of presently existing neurosecretory neurons have a
common evolutionary origin. Gene duplication, modification, and
cleavage of ancestral proteinaceous molecules are presumed to
have been involved in the development of chemical entities with
more and more diversified functional properties (see Wallis,
1975).

This process seems to be reflected by the fact that enzymatic
dissociation is responsible for the biosynthesis of most, if not all,
biologically active peptides known today (Tager and Steiner,
1974; Acher, 1976b). For example, the active nonapeptides stored
in the mammalian posterior lobe and the corresponding carrier
proteins are apparently not synthesized as such in the perikarya of
the respective hypothalamic neurons but are cleaved from a pre-
cursor of higher molecular weight (Sachs et al., 1969; Gainer,
Same and Brownstein, 1977). The fact that both components
make a strikingly sudden and simultaneous appearance early dur-
ing fetal development (Pearson, Goodman and Sachs, 1975) sup-
ports the view that they share the same macromolecular pre-
cursor. Moreover, the impressive structural similarity throughout
the entire vertebrate series of neurohypophysial hormones (Heller,
1974; Carraway and Leeman, 1975; Wallis, 1975; Acher, 1976a)

as well as their corresponding neurophysins (Capra and Walter,
1975; Acher. 1976b; Zimmerman, 1976) suggests that their pres-
ent preeursor moleeules (prohormones) are derived from closely
related ancestral proteins.

The same type of lineage can be claimed for hypophysiotropic
factors, the amino acid sequences of which are contained in
parent compounds of higher molecular weight. For example,
nonapeptides with hormonal activities of their own can play the
role of precursor for short-chain principles, such as the tripeptide
MIF (MSH-release inhibiting factor, melanostatin), a neurohor-
mone with different functional capacities (Walter, 1974; Reith et
al., 1977). Parenthetically, another unexpected feature about
nonapeptides is the recently reported presence of vasopressin,
unaccompanied by neurophysin, in a cell line from a human lung
carcinoma (Pettengill et al., 1977).

Information on analogous proteinaceous compounds in inverte-
brates is still sporadic. Nevertheless, histochemical and biochemi-
cal parallelisms can be recognized. For example, a chromato-
phorotropin that was chemically identified in crustaceans shows a
close resemblance to some of the small hypophysiotropic peptides
of mammals (Fernlund and Josefsson, 1972; Carlsen, Christensen
and Josefsson, 1976). Furthermore, a chemically synthesized oc-
tapeptide was shown to elicit pigment concentration in two types
of crustacean chromatophores in vitro and in vivo (Josefsson,
1975). A similar fully identified neuropeptide is the adipokinetic
hormone of insects (Stone et al., 1976). Another case in point is
the recent demonstration of immunoreactive TRF (thyrotropin
releasing factor) in the ganglia of some gastropods (Grimm-
J0rgensen, McKelvy and Jackson, 1975) where, for obvious rea-
sons, its function could resemble only the extrahypothalamic
activities demonstrated in vertebrates. An indication of the occur-
rence of such nonconventional interneuronal communication is
the recent observation (Takeuchi, Matsumoto and Mori, 1977)
that certain neurons of the snail Achatina are differentially af-
fected by fragments of some enzymatically treated nonapeptides,
e.g., oxytocin and vasotocin.

Finally, are there common denominators in the biosynthetic
and functional features of classical neurosecretory materials and

of other biologically active peptides produced by neurons and/or
glandular elements derived from neuroectodermal precursors (an-
terior pituitary and other members of the APUD cell series,
Pearse, 1976; Pearse and Takor Takor, 1976)?

In underscoring the neuroembryological and cytochemical fea-
tures shared by these cells, Pearse' s intriguing concept clarifies
the multiple occurrence, both within and outside the adult ner-
vous system, of a variety of regulatory peptides, including Sub-
stance P, enkephalins, endorphins, and several hypophysial
hormones. Therefore, all of these peptides may also have in
common the mode of their molecular evolution.

An example in support of this proposition is the mounting
evidence for the derivation of several such peptides with distinc-
tive physiological properties from a larger parent molecule, the
formerly enigmatic pituitary hormone /3-lipotropin (Li, 1964).
Among its subunits the endorphins (Goldstein, 1976; Guillemin,
1977) are currently receiving much attention because of their
analgesic and behavioral effects. The N-terminal of a-endorphin,
a short sequence (amino acid residues 61 through 65) apparently
representing the analgesically active core, precisely matches that
of methionine enkephalin, one of the two specific neuronal pen-
tapeptides already referred to (Cox, Goldstein and Li, 1976;
Guillemin, 1977; Guillemin, Ling and Burgus, 1976). Therefore,
these endogenous opiates could be generated from the prohor-
mone /3-lipotropin, either within the brain or in the pituitary, in
which case they could reach their sites of action via the circula-
tion or the cerebrospinal fluid (Reith et al., 1977).

There are also several indications of functional parallelisms.
One is the demonstration that, in company with several hypo-
physiotropic polypeptides, enkephalin and endorphin (Dupont et
al., 1977; Simantov and Snyder, 1977) as well as Substance P
(Kato et al., 1976) elicit the release of growth hormone, ACTH,
FSH, and prolactin. Another is that Substance P, like enkephalin,
endorphin, and several hypophysiotropins, may act as a modula-
tor of neuronal activity (see Zetler, 1976).

Neither enkephalins nor opiate receptors have thus far been
found among invertebrates (Snyder, 1977). Nevertheless, the fea-
tures that all the biologically active peptides known to date have

in common add up to the following generalization. In the course
o\' a long evolutionary history, ancestral macromolecular proteins
have given rise to a variety of related compounds. Multiple sites
of cleavage and molecular modulation seem to have resulted in
the acquisition and dissociation of diverse, e.g., hormonal, neu-
rotransmitter-! ike, and carrier functions, whereby one active prin-
ciple may act in more than one capacity. These possibilities,
borne out by phylogenetic and ontogenetic considerations, illus-
trate the principle of biochemical economy.

CONCLUSIONS

An examination of the evolutionary history of neural tissue,
and its pluripotentiality in primitive animals, points up its special
glandular attributes. The old inherited capacity for secretory ac-
tivity seems to have been put to use in multiple and specialized
ways at consecutive levels of the evolutionary scale. Therefore,
the spectrum of available neurochemical mediators and of modes
of information transfer that digress from standard synaptic trans-
mission is more diversified than previously assumed, even in
higher forms. Hormones derived from neural elements have re-
mained indispensable even after the appearance of the endocrine
system proper. The most unorthodox neurosecretory cells giving
rise to these blood-borne proteinaceous messengers, as well as
those signaling at closer range, have found their place within the
range of existing variants. Now that the versatility of the hypo-
thalamic neurosecretory centers and their major role in neuro-
endocrine integration have been clarified, current interest can turn
to extrahypothalamic neuropeptides. This nonconventional minor-
ity seems to have a relatively wide distribution within and outside
the central nervous system and to function in remarkable ways.
Further exploration of the glandular aspects of neuronal function
and their relationship with peptide-hormone-producing cells of
neuroectodermal origin holds much promise.

LITERATURE CITED

Acher, R.

1976a. Molecular evolution of the polypeptide hormones. In Polypeptide Hormones:
molecular and cellular aspects. Ciba Found. Symp. Ser. 41 (new series).
North Holland, Elsevier — Excerpta Medica, pp. 31-59.

1976b. Les neurophysines. Aspects moleculaires et cellulaires. Biochimie, vol. 58,
pp. 895-911.

Bargmann, W., and E. Scharrer

1951. The site of origin of the hormones of the posterior pituitary. Amer. Sci., vol.
39, pp. 255-259.

Barker, J.L. , and H. Gainer

1974. Peptide regulation of bursting pacemaker activity in a molluscan neurosecre-
tory cell. Science, vol. 184, pp. 1371-1373.

Brown, M., and W. Vale

1975. Central nervous system effects of hypothalamic peptides. Endocrinology, vol.
96, pp. 1333-1336.

Brownstein, M.J., M. Palkovits, J.M. Saavedra, and J.S. Kizer

1976. Distribution of hypothalamic hormones and neurotransmitters within the dien-
cephalon. In Martini, L. and W.F. Ganong (eds.), Frontiers in neuroen-
docrinology. New York, Raven Press, vol. 4, pp. 1-23.

Burnett. A.L. , and N.A. Diehl

1964. The nervous system of Hydra. III. The initiation of sexuality with special
reference to the nervous system. Jour. Exp. Zool., vol. 157, pp. 237-249.

Capra, J.D., and R. Walter

1975. Primary structure and evolution of neurophysins. Ann. New York Acad. Sci.,
vol. 248, pp. 397-407.

Carlsen, J., M. Christensen, and L. Josefsson

1976. Purification and chemical structure of the red pigment-concentrating hormone
of the prawn Leander adspersus . Gen. Comp. Endocrinol., vol. 30, pp.
327-331.

Carraway, R., and S.E. Leeman

1975. The amino acid sequence of a hypothalamic peptide, neurotensin. Jour. Biol.
Chem., vol. 250, pp. 1907-1911.

Constantinidis, J., F. Greissbuhler, J.M. Gaillard, T. Hovaguimian, and R. Tissot
1974. Enhancement of cerebral noradrenaline turnover by thyrotropin- releasing hor-
mone: evidence by fluorescence histochemistry. Experientia, vol. 30, pp.
1182-1183.

Cox. B.M.. A Goldstein, and C.H Li

1976. Opioid activity of a peptide. /3-lipotropin-<61-91). derived from /3-lipotropm
Ptac. Natl. Acad. Sci. USA. vol. 73. pp. 1821-1823

Davis. L.E.

1974. Ultrastructural studies of the development of nerve*- in H\dru Amer. Zool..
vol. 14. pp 551-573.

Dupont. A.. L. Cusan. F. Labrie. D.H. Co> . and C.H Li

1977. Stimulation of prolactin release in the rat by intraventricular injection of /3-
endorphin and methionine-enkephalin. Biochem. Biophys. Res. Comm.. vol.
75. pp. 76-82.

Femlund. P.. and L. Josefsson

1 9^2. Crustacean color-change hormone: amino acid sequence and chemical syn-
thesis. Science, vol. 177, pp. 173-175.

Gainer. H.. Y. Same, and M.J. Brownstein

1977. Neurophysin biosynthesis: conversion of a putative precursor during axonal
transport. Science, vol. 195. pp. 1354-1356.

Goldstein. A

1976. Opioid peptides (endorphins) in pituitary and brain. Science, vol. 193. pp.
1081-1086.

Grasso. M.. and M. Benazzi

1973. Genetic and physiologic control of fissioning and sexuality in planarians. Jour.
Embryol. Exp. Morph.. vol. 30. pp. 317-328.

Grimm-Jorgensen. Y.. J.F. McKelvy . and I.M.D. Jackson

1975. Immunoreactive thyTotrophin releasing factor in gastropod circumoesophageal
ganglia. Nature, vol. 254. p. 420.

Guillemin. R.

1977. Endorphins, brain peptides that act like opiates. New England Jour. Med..
vol. 296. pp. 226-228.

Guillemin. R.. N. Ling, and R. Burgus

1976. Endorphines. peptides, d'origine hypothalamique et neurohypophysaire a ac-
rivite morphinomimetique. Isolement et structure moleculaire de l'a-en-
dorphine. Compt. Rend. Acad. Sci.. Paris. Ser. D. vol. 282. pp. 783-785.

Hauenschild. C.

1974. Endokrine Beeinflussung der geschlechtlichen Entvucklung einiger Poly-
chaeten. Fortschritte der Zoologie. vol. 22, pp. 75-92. Stuttgart. Gustav
Fischer Verlag.

Heller. H

1974. Molecular aspects in comparative endocrinology. Gen. Comp. Endocrinol.,
vol. 22. pp. 315-332.

Highnam, K.C., and L. Hill
1977. The comparative endocrinology of the invertebrates. Second edition. London,
Edward Arnold, pp. 357.

Horst, W.D., and N. Spirt

1974. A possible mechanism for the anti-depressant activity of thyrotropin releasing
hormone. Life Sci., vol. 15, pp. 1073-1082.

Hughes, J., T. Smith, B. Morgan, and L. Fothergill

1975. Purification and properties of enkephalin — the possible endogenous ligand for
the morphine receptor. Life Sci., vol. 16, pp. 1753-1758.

Josefsson, L.

1975. Structure and function of crustacean chromatophorotropins. Gen. Comp. Endo-
crinol., vol. 25, pp. 199-202.

Kato, Y., K. Chihara, S. Ohgo, Y. Iwasaki, H. Abe, and H. Imura

1976. Growth hormone and prolactin release by substance P in rats. Life Sci., vol.
19, pp. 441-446.

Lentz, T.L.

1968. Primitive nervous systems. New Haven and London, Yale Univ. Press.

Li, C.H.

1964. Lipotropin, a new active peptide from pituitary glands. Nature, vol. 201, p.
924.

Lote, C.J., J. P. Gent, J.H. Wolstencroft, and M. Szelke

1976. An inhibitory tripeptide from cat spinal cord. Nature, vol. 264, pp. 188-189.

McCann, S.M., and R.L. Moss

1975. Putative neurotransmitters involved in discharging gonadotropin-releasing
neurohormones and the action of LH-releasing hormone on the CNS. Life
Sci., vol. 16, pp. 833-852.

McNeill, T.H., G.P. Kozlowski, J.H. Abel, Jr., and E.A. Zimmerman

1976. Neurosecretory pathways in the mallard duck (Anas platyrhynchos) brain:
localization by aldehyde fuchsin and immunoperoxidase techniques for
neurophysin (NP) and gonadotrophin releasing hormone (Gn-RH). Endocrinol-
ogy, vol. 99, pp. 1323-1332.

Nicoll, R.A.

1977. Excitatory action of TRH on spinal motoneurones . Nature, vol. 265, pp.
242-243.

Pa vans de Ceccatty, M.
1966. Ultrastructures et rapports des cellules mesenchymateuses de type nerveux de
l'Eponge Tethya lyncurium Lmk. Ann. Sci. Nat. Zool., vol. 8, pp. 577-614.
1974. The origin of the integrative systems: a change in view derived from research
on coelenterates and sponges. Perspect. Biol. Med., vol. 17, pp. 379-390.

Pearse. AGE.

1976. Peptides m brain and intestine. Nature, vol. 262. pp. 92-94
Please, AGE. and T Takor Takor

1976. Neuroendocrine embryolog) and the APUD concept. Clin. Endocrinol., vol.
5, :: i k-244v

Pearson. D.B.. R. Goodman, and H. Sachs

1975. Stimulated vasopressin synthesis b> a fetal h\pothalamic factor. Science, vol.
187. pp. 1081-1082.

Pettengill. OS.. C.S. Faulkner. D.H. Wurster-Hill. L.H. Maurer. G.D. Sorenson. A.G.
Robinson, and E.A. Zimmerman

1977. Isolation and characterization of a hormone-producing cell line from human
small cell anaplastic carcinoma of the lung. Jour. Natl. Cancer Inst., vol. 58.
pp. 511-518.

Pilgrim. C.

1974. Histochemical differentiation of hypothalamic areas. Progr. Brain Res., vol.
41. pp. 97-110.

Plotnikoff. N.P.. W.F. White. A.J. Kastin. and A.V. Schally

1975. Gonadotropin releasing hormone (GnRH): neuropharmacological studies. Life
Sci.. vol. 17. pp. 1685-1692.

Prange, A.J.. Jr.. G.R. Breese. I.C. Wilson, and M.A. Lipton

1975a. Brain-behavioral effects of hypothalamic releasing hormones. Anat. Neuroen-
docrinol., Int. Conf. Neurobiology of CNS-Hormone Interactions. Chapel
Hill. 1974. pp. 357-366.

1975b. Behavioral effects of hypothalamic releasing hormones in animals and men. In
Gispen. W.H.. T.B. van Wimersma Greidanus. B. Bohus and D. de Wied
(eds.). Progress in brain research. Amsterdam, Oxford, New York, Elsevier.
vol. 42. pp. 1-9.

Reith. ME. A.. P. Schotman. W.H. Gispen. and D. de Wied

1977. Pituitary peptides as modulators of neural functioning. Trends in Biochem.
Sci.. vol. 2. pp. 56-58.

Rodriguez. E.M.

1976. The cerebrospinal fluid as a pathway in neuroendocrine integration. Jour.
Endocrinol., vol. 71. pp. 407-443.

Sachs. H.

1969. Neurosecretion. Adv. Enzymol.. vol. 32. pp. 327-372.

Sachs. H.. P. Fawcett. Y. Takabatake. and R. Portanova

1969. Biosynthesis and release of vasopressin and neurophysin. Recent Progr. Hor-
mone Res.. 25: 447-491.

Scharrer, B.

1970. General principles of neuroendocrine communication. In Schmitt, F.O. (ed.),
The neurosciences: second study program. New York, The Rockefeller Univ.
Press, pp. 519-529.

1972. Neuroendocrine communication (neurohormonal, neurohumoral, and interme-
diate). In Ariens Kappers, J., and J. P. Schade (eds.), Progress in brain
research. Amsterdam, London, New York, Elsevier, vol. 38, pp. 7-18.

1974. The spectrum of neuroendocrine communication. In Recent studies of hypo-
thalamic function. Int. Symp. Calgary, 1973. Basel, Karger, pp. 8-16.

1976. Neurosecretion — comparative and evolutionary aspects. In Corner, M.A., and
D.F. Swaab (eds.), Perspectives in brain research. Progress in brain research.
Amsterdam, London, New York, Elsevier/North Holland Biomedical Press,
vol. 45, pp. 125-137.

Scharrer, E.

1928. Die Lichtempfindlichkeit blinder Elritzen (Untersuchungen iiber das
Zwischenhirn der Fische. I.). Z. Vgl. Physiol., vol. 7, pp. 1-38.

Schwartz, I.L., and R. Walter

1974. Neurohypophyseal hormones as precursors of hypophysiotropic hormones. Is-
rael Jour. Med. Sci.. vol. 10, pp. 1288-1293.

Simantov, R., and S.H. Snyder

1977. Opiate receptor binding in the pituitary gland. Brain Res., vol. 124, pp.
178-184.

Smith, B.H., and G.W. Kreutzberg (eds.)

1976. Neuron-target cell interactions. Neurosciences Res. Prog. Bull., vol. 14, pp.
212-453.

Snyder, S.H.

1977. Opiate receptors and internal opiates. Sci. Amer., vol. 236, pp. 44-56.
Speidel, C.C.

1919. Gland-cells of internal secretion in the spinal cord of the skates. Carnegie Inst.
Washington Publ. No. 13, pp. 1-31.

Sterba, G.

1974. Ascending neurosecretory pathways of the peptidergic type. In F. Knowles,
and L. Vollrath (eds.), Neurosecretion — the final neuroendocrine pathway. VI
Internatl. Symp. on Neurosecretion, London, 1973. Berlin, Heidelberg, New
York, Springer- Verlag, pp. 38-47.

Stone, J. V., W. Mordue, K.E. Batley, and H.R. Morris

1976. Structure of locust adipokinetic hormone, a neurohormone that regulates lipid
utilisation during flight. Nature, vol. 263, pp. 207-211.

Takeuchi, H., M. Matsumoto, and A. Mori

1977. Modification of effects of biologically active peptides, caused by enzyme

treatment, on the excitability of identifiable uiant neurones of an African giant

snail iAchulina fulica Feruss.ie) l.\penenli.i. vol. 33, pp. 249-251.
Tager, H.S.. and D.F. Steiner

1974. Peptide hormones. Ann. Rev. Bioehem.. vol. 43. pp. 509-538.
Ude. J.

1976. Immunclektronenmikroskopische Ultradiinnschnittmarkierung-Ergebnisse und
Probleme. Acta Histochem., Suppl. XVII. pp. 271-283.

Vineent. J.D., and E. Amauld

1975. Vasopressin as a neurotransmitter in the eentral nervous system: some evi-
dence from the supraoptic neurosecretory system. In Gispen. W.H., T.B. van
Wimersma Greidanus, B. Bohus, and D. de Wied (eds.). Progress in brain
research. Amsterdam, Oxford. New York, Elsevier, vol. 42, pp. 57-66.

Wall is. M.

1975. The molecular evolution of pituitary hormones. Biol. Rev., vol. 50, pp.
35-98.

Walter. R.

1974. Oxytocin and other peptide hormones as prohormones. In Hatotani, N. (ed.),
Psychoneuroendocrinology. Basel, Karger, pp. 285-294.

1975. Neurophysins: carriers of peptide hormones. Ann. N.Y. Acad. Sci., vol. 248.
pp. 512.

Watkins. W.B.

1975. Immunohistochemical demonstration of neurophysin in the hypo-
thalamoneurohypophysial system. Internat. Rev. Cytol., vol. 41, pp. 241-284.

Waziri. R.

1975. Effects of thyrotropin releasing hormone on cholinergic synaptic transmission
in Aplysia. Neuroscience Abstracts, Soc. for Neuroscience 5th Annual Meet-
ing, p. 447 (Abstr. #694).

de Wied. D.. T.B. van Wimersma Greidanus, B. Bohus. I. Urban, and W.H. Gispen

1976. Vasopressin and memory consolidation. In Corner, M.A. and D.F. Swaab
(eds.). Perspectives in brain research. Progress in brain research. Amsterdam,
London, New York, Elsevier/North Holland Biomedical Press, vol. 45. pp.
181-194.

Zetler. G.

1976. The peptidergic neuron: a working hypothesis. Bioehem. Pharmacol., vol. 25,
pp. 1817-1818.

Zimmerman, E.A.

1976. Localization of hypothalamic hormones by immunocytochemical techniques.
In Martini. L. and W.F. Ganong (eds.). Frontiers in neuroendocrinology. New
York, Raven Press, vol. 4, pp. 25-62.